| base_model: aaditya/phi_2 |
| model_type: AutoModelForCausalLM |
| tokenizer_type: AutoTokenizer |
|
|
| load_in_8bit: false |
| load_in_4bit: true |
| strict: false |
|
|
|
|
| datasets: |
| - path: aaditya/alpaca_subset_1 |
| type: alpaca |
| val_set_size: 0 |
| output_dir: . |
|
|
| sequence_len: 2048 |
| sample_packing: true |
| pad_to_sequence_len: true |
|
|
| adapter: qlora |
| lora_model_dir: |
| lora_r: 64 |
| lora_alpha: 32 |
| lora_dropout: 0.05 |
| lora_target_linear: true |
| lora_fan_in_fan_out: |
|
|
| wandb_project: |
| wandb_entity: |
| wandb_watch: |
| wandb_name: |
| wandb_log_model: |
|
|
| gradient_accumulation_steps: 2 |
| micro_batch_size: 20 |
| num_epochs: 1 |
| optimizer: adamw_torch |
| adam_beta2: 0.95 |
| adam_epsilon: 0.00001 |
| max_grad_norm: 1.0 |
| lr_scheduler: cosine |
| learning_rate: 0.000003 |
|
|
| train_on_inputs: false |
| group_by_length: false |
| bf16: auto |
| fp16: |
| tf32: true |
|
|
| gradient_checkpointing: true |
| gradient_checkpointing_kwargs: |
| use_reentrant: True |
| early_stopping_patience: |
| resume_from_checkpoint: |
| local_rank: |
| logging_steps: 1 |
| xformers_attention: |
| flash_attention: true |
|
|
| warmup_steps: 100 |
| evals_per_epoch: 4 |
| saves_per_epoch: 1 |
| debug: |
| deepspeed: |
| weight_decay: 0.1 |
| fsdp: |
| fsdp_config: |
| resize_token_embeddings_to_32x: true |
| special_tokens: |
| pad_token: "<|endoftext|>" |
