Prummn commited on
Commit
c9c6229
·
verified ·
1 Parent(s): fd04e67

Add files using upload-large-folder tool

Browse files
Files changed (50) hide show
  1. checkpoint-1000/README.md +202 -0
  2. checkpoint-1000/adapter_config.json +26 -0
  3. checkpoint-1000/adapter_model.safetensors +3 -0
  4. checkpoint-1000/additional_config.json +1 -0
  5. checkpoint-1000/args.json +353 -0
  6. checkpoint-1000/trainer_state.json +0 -0
  7. checkpoint-2000/README.md +202 -0
  8. checkpoint-2000/adapter_config.json +26 -0
  9. checkpoint-2000/additional_config.json +1 -0
  10. checkpoint-2000/args.json +353 -0
  11. checkpoint-2000/optimizer.pt +3 -0
  12. checkpoint-2000/trainer_state.json +0 -0
  13. checkpoint-3000/README.md +202 -0
  14. checkpoint-3000/adapter_config.json +26 -0
  15. checkpoint-3000/additional_config.json +1 -0
  16. checkpoint-3000/args.json +353 -0
  17. checkpoint-3000/rng_state_1.pth +3 -0
  18. checkpoint-3000/rng_state_3.pth +3 -0
  19. checkpoint-3000/rng_state_4.pth +3 -0
  20. checkpoint-3000/rng_state_6.pth +3 -0
  21. checkpoint-3000/scheduler.pt +3 -0
  22. checkpoint-3000/trainer_state.json +0 -0
  23. checkpoint-3000/training_args.bin +3 -0
  24. checkpoint-4000/README.md +202 -0
  25. checkpoint-4000/adapter_config.json +26 -0
  26. checkpoint-4000/additional_config.json +1 -0
  27. checkpoint-4000/args.json +353 -0
  28. checkpoint-4000/trainer_state.json +0 -0
  29. checkpoint-5000/README.md +202 -0
  30. checkpoint-5000/adapter_config.json +26 -0
  31. checkpoint-5000/additional_config.json +1 -0
  32. checkpoint-5000/args.json +353 -0
  33. checkpoint-5000/trainer_state.json +0 -0
  34. checkpoint-6000/README.md +202 -0
  35. checkpoint-6000/adapter_config.json +26 -0
  36. checkpoint-6000/additional_config.json +1 -0
  37. checkpoint-6000/args.json +353 -0
  38. checkpoint-6000/optimizer.pt +3 -0
  39. checkpoint-6000/trainer_state.json +0 -0
  40. checkpoint-7000/README.md +202 -0
  41. checkpoint-7000/adapter_config.json +26 -0
  42. checkpoint-7000/adapter_model.safetensors +3 -0
  43. checkpoint-7000/additional_config.json +1 -0
  44. checkpoint-7000/args.json +353 -0
  45. checkpoint-7000/trainer_state.json +0 -0
  46. checkpoint-8000/README.md +202 -0
  47. checkpoint-8000/adapter_config.json +26 -0
  48. checkpoint-8000/additional_config.json +1 -0
  49. checkpoint-8000/args.json +353 -0
  50. checkpoint-8000/trainer_state.json +0 -0
checkpoint-1000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-1000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-1000/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb4f4c765d179f4c8e7dbd0b5b9e0ddc9ee6094e99b3266b4e0f18809bee8022
3
+ size 11988912
checkpoint-1000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-1000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-1000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-2000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-2000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-2000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa150451a3814972fd0a01608e7c2fd9f5b2194cf87a7f013ec39fe8330bb068
3
+ size 24205999
checkpoint-2000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-3000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-3000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-3000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-3000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8499843c8b47b8d3e26e7df6714f3e727463d45e7c647a48a1515365720a2b98
3
+ size 16389
checkpoint-3000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e692434f912136e27f627d369442b038ba26a4828292d7804a35e5aa8001347c
3
+ size 16389
checkpoint-3000/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad9d8852e12491d911950433bc83c305c061b2958ddd8399334ba407fda77789
3
+ size 16389
checkpoint-3000/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67b78e49976b12a9145196d745166b4b5d486e393f415d7762000c1afda1c5c1
3
+ size 16389
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b5674374a3c68e3dd9852cd7d2e14d635bcd5a1495192e90ec1a1995d5b8b77
3
+ size 1465
checkpoint-3000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d571ea992b7978e0e3bb25e34f2ce8c1da8261ea4665c7b52d47e23a136ed71f
3
+ size 7249
checkpoint-4000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-4000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-4000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-4000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-4000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-5000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-5000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-5000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-5000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-5000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-6000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-6000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-6000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-6000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-6000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c12ec1683d77a430fa795525c7f95ff790d81e0456d044f840e75500cb972a38
3
+ size 24205999
checkpoint-6000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-7000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-7000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-7000/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90f09dbdda1bf889fab3f54230cd713e3c827d847967eaea9a3d644f59aa792f
3
+ size 11988912
checkpoint-7000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-7000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-7000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-8000/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: /data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.11.1
checkpoint-8000/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": [],
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": "^(encoder.*\\.(value|mlp.0|mlp.2|out|query|key)|adapter.*\\.(linear1|linear2))$",
23
+ "task_type": "CAUSAL_LM",
24
+ "use_dora": false,
25
+ "use_rslora": false
26
+ }
checkpoint-8000/additional_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"lora_dtype": null, "lorap_lr_ratio": null, "lorap_emb_lr": 1e-06}
checkpoint-8000/args.json ADDED
@@ -0,0 +1,353 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "output_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801",
3
+ "overwrite_output_dir": false,
4
+ "do_train": false,
5
+ "do_eval": false,
6
+ "do_predict": false,
7
+ "eval_strategy": "steps",
8
+ "prediction_loss_only": false,
9
+ "per_device_train_batch_size": 8,
10
+ "per_device_eval_batch_size": 8,
11
+ "per_gpu_train_batch_size": null,
12
+ "per_gpu_eval_batch_size": null,
13
+ "gradient_accumulation_steps": 2,
14
+ "eval_accumulation_steps": null,
15
+ "eval_delay": 0,
16
+ "torch_empty_cache_steps": null,
17
+ "learning_rate": 1e-06,
18
+ "weight_decay": 0.01,
19
+ "adam_beta1": 0.9,
20
+ "adam_beta2": 0.95,
21
+ "adam_epsilon": 1e-08,
22
+ "max_grad_norm": 2.5,
23
+ "num_train_epochs": 2.0,
24
+ "max_steps": -1,
25
+ "lr_scheduler_type": "cosine",
26
+ "lr_scheduler_kwargs": null,
27
+ "warmup_ratio": 0.02,
28
+ "warmup_steps": 0,
29
+ "log_level": "passive",
30
+ "log_level_replica": "warning",
31
+ "log_on_each_node": true,
32
+ "logging_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs",
33
+ "logging_strategy": "steps",
34
+ "logging_first_step": true,
35
+ "logging_steps": 1,
36
+ "logging_nan_inf_filter": true,
37
+ "save_strategy": "steps",
38
+ "save_steps": 500.0,
39
+ "save_total_limit": 40,
40
+ "save_safetensors": true,
41
+ "save_on_each_node": false,
42
+ "save_only_model": false,
43
+ "restore_callback_states_from_checkpoint": false,
44
+ "no_cuda": false,
45
+ "use_cpu": false,
46
+ "use_mps_device": false,
47
+ "seed": 42,
48
+ "data_seed": 42,
49
+ "jit_mode_eval": false,
50
+ "use_ipex": false,
51
+ "bf16": true,
52
+ "fp16": false,
53
+ "fp16_opt_level": "O1",
54
+ "half_precision_backend": "auto",
55
+ "bf16_full_eval": false,
56
+ "fp16_full_eval": false,
57
+ "tf32": null,
58
+ "local_rank": 0,
59
+ "ddp_backend": null,
60
+ "tpu_num_cores": null,
61
+ "tpu_metrics_debug": false,
62
+ "debug": null,
63
+ "dataloader_drop_last": false,
64
+ "eval_steps": 200.0,
65
+ "dataloader_num_workers": null,
66
+ "dataloader_prefetch_factor": null,
67
+ "past_index": -1,
68
+ "run_name": "step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all",
69
+ "disable_tqdm": null,
70
+ "remove_unused_columns": true,
71
+ "label_names": null,
72
+ "load_best_model_at_end": false,
73
+ "metric_for_best_model": "loss",
74
+ "greater_is_better": false,
75
+ "ignore_data_skip": true,
76
+ "fsdp": "",
77
+ "fsdp_min_num_params": 0,
78
+ "fsdp_config": null,
79
+ "fsdp_transformer_layer_cls_to_wrap": null,
80
+ "accelerator_config": {
81
+ "dispatch_batches": false
82
+ },
83
+ "deepspeed": null,
84
+ "label_smoothing_factor": 0.0,
85
+ "optim": "adamw_torch",
86
+ "optim_args": null,
87
+ "adafactor": false,
88
+ "group_by_length": false,
89
+ "length_column_name": "length",
90
+ "report_to": [
91
+ "wandb"
92
+ ],
93
+ "ddp_find_unused_parameters": null,
94
+ "ddp_bucket_cap_mb": null,
95
+ "ddp_broadcast_buffers": null,
96
+ "dataloader_pin_memory": true,
97
+ "dataloader_persistent_workers": false,
98
+ "skip_memory_metrics": true,
99
+ "use_legacy_prediction_loop": false,
100
+ "push_to_hub": false,
101
+ "resume_from_checkpoint": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
102
+ "hub_model_id": null,
103
+ "hub_strategy": "every_save",
104
+ "hub_token": null,
105
+ "hub_private_repo": null,
106
+ "hub_always_push": false,
107
+ "gradient_checkpointing": true,
108
+ "gradient_checkpointing_kwargs": "{\"use_reentrant\": false}",
109
+ "include_inputs_for_metrics": false,
110
+ "include_for_metrics": [],
111
+ "eval_do_concat_batches": true,
112
+ "fp16_backend": "auto",
113
+ "evaluation_strategy": "steps",
114
+ "push_to_hub_model_id": null,
115
+ "push_to_hub_organization": null,
116
+ "push_to_hub_token": null,
117
+ "mp_parameters": "",
118
+ "auto_find_batch_size": false,
119
+ "full_determinism": false,
120
+ "torchdynamo": null,
121
+ "ray_scope": "last",
122
+ "ddp_timeout": 18000000,
123
+ "torch_compile": false,
124
+ "torch_compile_backend": null,
125
+ "torch_compile_mode": null,
126
+ "dispatch_batches": null,
127
+ "split_batches": null,
128
+ "include_tokens_per_second": false,
129
+ "include_num_input_tokens_seen": false,
130
+ "neftune_noise_alpha": null,
131
+ "optim_target_modules": null,
132
+ "batch_eval_metrics": false,
133
+ "eval_on_start": false,
134
+ "use_liger_kernel": false,
135
+ "eval_use_gather_object": false,
136
+ "average_tokens_across_devices": false,
137
+ "sortish_sampler": false,
138
+ "predict_with_generate": false,
139
+ "generation_max_length": null,
140
+ "generation_num_beams": null,
141
+ "generation_config": null,
142
+ "tuner_backend": "peft",
143
+ "vit_gradient_checkpointing": null,
144
+ "router_aux_loss_coef": 0.0,
145
+ "enable_dft_loss": false,
146
+ "enable_channel_loss": false,
147
+ "check_model": true,
148
+ "acc_strategy": "token",
149
+ "train_dataloader_shuffle": true,
150
+ "max_epochs": null,
151
+ "aligner_lr": 2e-06,
152
+ "vit_lr": 1e-06,
153
+ "use_logits_to_keep": null,
154
+ "ds3_gather_for_generation": true,
155
+ "resume_only_model": true,
156
+ "optimizer": null,
157
+ "loss_type": null,
158
+ "metric": null,
159
+ "eval_use_evalscope": false,
160
+ "eval_dataset": [],
161
+ "eval_dataset_args": null,
162
+ "eval_limit": null,
163
+ "eval_generation_config": null,
164
+ "extra_eval_args": null,
165
+ "use_flash_ckpt": false,
166
+ "use_ray": false,
167
+ "ray_exp_name": null,
168
+ "device_groups": null,
169
+ "model": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
170
+ "model_type": "step_audio2_mini",
171
+ "model_revision": null,
172
+ "task_type": "causal_lm",
173
+ "torch_dtype": "bfloat16",
174
+ "attn_impl": null,
175
+ "new_special_tokens": [],
176
+ "num_labels": null,
177
+ "problem_type": null,
178
+ "rope_scaling": null,
179
+ "device_map": null,
180
+ "max_memory": {},
181
+ "max_model_len": null,
182
+ "local_repo_path": null,
183
+ "init_strategy": null,
184
+ "template": "step_audio2_mini",
185
+ "system": null,
186
+ "max_length": 16384,
187
+ "truncation_strategy": "delete",
188
+ "max_pixels": null,
189
+ "agent_template": null,
190
+ "norm_bbox": null,
191
+ "use_chat_template": true,
192
+ "padding_free": false,
193
+ "padding_side": "right",
194
+ "loss_scale": "default",
195
+ "sequence_parallel_size": 1,
196
+ "response_prefix": null,
197
+ "template_backend": "swift",
198
+ "dataset": [
199
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_train90.jsonl"
200
+ ],
201
+ "val_dataset": [
202
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
203
+ ],
204
+ "cached_dataset": [],
205
+ "cached_val_dataset": [],
206
+ "split_dataset_ratio": 0.0,
207
+ "dataset_num_proc": 1,
208
+ "load_from_cache_file": false,
209
+ "dataset_shuffle": true,
210
+ "val_dataset_shuffle": false,
211
+ "streaming": false,
212
+ "interleave_prob": null,
213
+ "stopping_strategy": "first_exhausted",
214
+ "shuffle_buffer_size": 1000,
215
+ "download_mode": "reuse_dataset_if_exists",
216
+ "columns": {},
217
+ "strict": false,
218
+ "model_name": null,
219
+ "model_author": null,
220
+ "custom_dataset_info": [],
221
+ "quant_method": null,
222
+ "quant_bits": null,
223
+ "hqq_axis": null,
224
+ "bnb_4bit_compute_dtype": "bfloat16",
225
+ "bnb_4bit_quant_type": "nf4",
226
+ "bnb_4bit_use_double_quant": true,
227
+ "bnb_4bit_quant_storage": null,
228
+ "max_new_tokens": 64,
229
+ "temperature": 0.0,
230
+ "top_k": null,
231
+ "top_p": null,
232
+ "repetition_penalty": null,
233
+ "num_beams": 1,
234
+ "stream": false,
235
+ "stop_words": [],
236
+ "logprobs": false,
237
+ "top_logprobs": null,
238
+ "ckpt_dir": "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000",
239
+ "lora_modules": [],
240
+ "train_type": "lora",
241
+ "adapters": [
242
+ "/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000"
243
+ ],
244
+ "external_plugins": [],
245
+ "model_kwargs": {},
246
+ "load_args": false,
247
+ "load_data_args": false,
248
+ "packing": false,
249
+ "packing_length": null,
250
+ "packing_num_proc": 1,
251
+ "lazy_tokenize": true,
252
+ "custom_register_path": [],
253
+ "use_hf": false,
254
+ "ignore_args_error": false,
255
+ "use_swift_lora": false,
256
+ "freeze_parameters": [],
257
+ "freeze_parameters_regex": null,
258
+ "freeze_parameters_ratio": 0.0,
259
+ "trainable_parameters": [],
260
+ "trainable_parameters_regex": null,
261
+ "freeze_llm": false,
262
+ "freeze_vit": false,
263
+ "freeze_aligner": false,
264
+ "target_modules": [
265
+ "all-linear"
266
+ ],
267
+ "target_regex": null,
268
+ "target_parameters": null,
269
+ "modules_to_save": [],
270
+ "lora_rank": 8,
271
+ "lora_alpha": 32,
272
+ "lora_dropout": 0.1,
273
+ "lora_bias": "none",
274
+ "lora_dtype": null,
275
+ "lorap_lr_ratio": null,
276
+ "use_rslora": false,
277
+ "use_dora": false,
278
+ "lora_ga_batch_size": 2,
279
+ "lora_ga_iters": 2,
280
+ "lora_ga_max_length": 1024,
281
+ "lora_ga_direction": "ArB2r",
282
+ "lora_ga_scale": "stable",
283
+ "lora_ga_stable_gamma": 16,
284
+ "init_weights": true,
285
+ "fourier_n_frequency": 2000,
286
+ "fourier_scaling": 300.0,
287
+ "boft_block_size": 4,
288
+ "boft_block_num": 0,
289
+ "boft_n_butterfly_factor": 1,
290
+ "boft_dropout": 0.0,
291
+ "vera_rank": 256,
292
+ "vera_projection_prng_key": 0,
293
+ "vera_dropout": 0.0,
294
+ "vera_d_initial": 0.1,
295
+ "adapter_act": "gelu",
296
+ "adapter_length": 128,
297
+ "use_galore": false,
298
+ "galore_target_modules": null,
299
+ "galore_rank": 128,
300
+ "galore_update_proj_gap": 50,
301
+ "galore_scale": 1.0,
302
+ "galore_proj_type": "std",
303
+ "galore_optim_per_parameter": false,
304
+ "galore_with_embedding": false,
305
+ "galore_quantization": false,
306
+ "galore_proj_quant": false,
307
+ "galore_proj_bits": 4,
308
+ "galore_proj_group_size": 256,
309
+ "galore_cos_threshold": 0.4,
310
+ "galore_gamma_proj": 2,
311
+ "galore_queue_size": 5,
312
+ "adalora_target_r": 8,
313
+ "adalora_init_r": 12,
314
+ "adalora_tinit": 0,
315
+ "adalora_tfinal": 0,
316
+ "adalora_deltaT": 1,
317
+ "adalora_beta1": 0.85,
318
+ "adalora_beta2": 0.85,
319
+ "adalora_orth_reg_weight": 0.5,
320
+ "llamapro_num_new_blocks": 4,
321
+ "llamapro_num_groups": null,
322
+ "lisa_activated_layers": 0,
323
+ "lisa_step_interval": 20,
324
+ "reft_layer_key": null,
325
+ "reft_layers": null,
326
+ "reft_rank": 4,
327
+ "reft_intervention_type": "LoreftIntervention",
328
+ "reft_args": null,
329
+ "swanlab_token": null,
330
+ "swanlab_project": null,
331
+ "swanlab_workspace": null,
332
+ "swanlab_exp_name": null,
333
+ "swanlab_lark_webhook_url": null,
334
+ "swanlab_lark_secret": null,
335
+ "swanlab_mode": "cloud",
336
+ "add_version": true,
337
+ "create_checkpoint_symlink": false,
338
+ "zero_hpz_partition_size": null,
339
+ "deepspeed_autotp_size": null,
340
+ "early_stop_interval": null,
341
+ "rank": 0,
342
+ "global_world_size": 8,
343
+ "local_world_size": 8,
344
+ "model_suffix": "Step-Audio-2-mini",
345
+ "model_info": "ModelInfo(model_type='step_audio2_mini', model_dir='/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini', torch_dtype=torch.bfloat16, max_model_len=16384, quant_method=None, quant_bits=None, rope_scaling=None, is_moe_model=False, is_multimodal=True, config=None, task_type='causal_lm', num_labels=None)",
346
+ "model_meta": "ModelMeta(model_type='step_audio2_mini', model_groups=[ModelGroup(models=[Model(ms_model_id='stepfun-ai/Step-Audio-2-mini', hf_model_id='stepfun-ai/Step-Audio-2-mini', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='step_audio2_mini', get_function=<function get_model_tokenizer_step_audio2_mini at 0x7efc2c8e24d0>, model_arch=MultiModelKeys(arch_name='step_audio2_mini', embedding=None, module_list=None, lm_head=None, q_proj=None, k_proj=None, v_proj=None, o_proj=None, attention=None, mlp=None, down_proj=None, qkv_proj=None, qk_proj=None, qa_proj=None, qb_proj=None, kv_proj=None, kva_proj=None, kvb_proj=None, language_model=['model', 'lm_head'], aligner=['adapter'], vision_tower=['encoder'], generator=[]), architectures=['StepAudio2ForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=True, is_reward=False, is_reranker=False, task_type=None, ignore_patterns=None, requires=['transformers==4.53.3', 'torchaudio', 'librosa'], tags=['audio'])",
347
+ "model_dir": "/data/liuzihang/haobin/Step-Audio2/Step-Audio-2-mini",
348
+ "_val_dataset_exists": [
349
+ "/data/liuzihang/haobin/pangkaiyu/sft/jsonl_sft/lora_merged_0130_55+5+5+6+4_drop_signal_val5.jsonl"
350
+ ],
351
+ "hub": "<class 'swift.hub.hub.MSHub'>",
352
+ "training_args": "Seq2SeqTrainingArguments(output_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=<IntervalStrategy.STEPS: 'steps'>, prediction_loss_only=False, per_device_train_batch_size=8, per_device_eval_batch_size=8, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=2, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=1e-06, weight_decay=0.01, adam_beta1=0.9, adam_beta2=0.95, adam_epsilon=1e-08, max_grad_norm=2.5, num_train_epochs=2.0, max_steps=-1, lr_scheduler_type=<SchedulerType.COSINE: 'cosine'>, lr_scheduler_kwargs=None, warmup_ratio=0.02, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align+llm-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF_2e6_all/v2-20260215-150801/runs', logging_strategy=<IntervalStrategy.STEPS: 'steps'>, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=<SaveStrategy.STEPS: 'steps'>, save_steps=500, save_total_limit=40, save_safetensors=True, save_on_each_node=False, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=200, dataloader_num_workers=1, dataloader_prefetch_factor=10, past_index=-1, run_name='step-audio2-mini-enc+align+llm-whole0130_signal_new_dpdc_lora-8gpu-bs16__gckF_2e6_all', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='loss', greater_is_better=False, ignore_data_skip=True, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed=None, label_smoothing_factor=0.0, optim=<OptimizerNames.ADAMW_TORCH: 'adamw_torch'>, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint='/data/liuzihang/haobin/pangkaiyu/output/output_step_audio2_mini-encoder+align-whole0130_signal_new1_dpdc-lora-1gpu-bs16_4_gckF/v0-20260130-120145/checkpoint-6000', hub_model_id=None, hub_strategy=<HubStrategy.EVERY_SAVE: 'every_save'>, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs={'use_reentrant': False}, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=18000000, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, sortish_sampler=False, predict_with_generate=False, generation_max_length=None, generation_num_beams=None, generation_config=None, tuner_backend='peft', vit_gradient_checkpointing=True, router_aux_loss_coef=0.0, enable_dft_loss=False, enable_channel_loss=False, check_model=True, acc_strategy='token', train_dataloader_shuffle=True, max_epochs=None, aligner_lr=2e-06, vit_lr=1e-06, use_logits_to_keep=None, ds3_gather_for_generation=True, resume_only_model=True, optimizer='multimodal', loss_type=None, metric=None, eval_use_evalscope=False, eval_dataset=[], eval_dataset_args=None, eval_limit=None, eval_generation_config=None, extra_eval_args=None, use_flash_ckpt=False, sft_alpha=0, chord_sft_dataset=[], chord_sft_per_device_train_batch_size=None, chord_enable_phi_function=False, chord_mu_warmup_steps=None, chord_mu_decay_steps=None, chord_mu_peak=None, chord_mu_valley=None, train_type='lora', local_repo_path=None, galore_config=None, padding_side='right', padding_free=False, task_type='causal_lm', problem_type=None)"
353
+ }
checkpoint-8000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff