Hugging Face's logo

Yuehao
/
llada

Safetensors
Model card Files
xet
 Community
Yuehao commited on
Commit
3181cc8
·
verified ·
1 Parent(s): 7af472a

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/added_tokens.json +3 -0
  2. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/chat_template.jinja +11 -0
  3. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/config.json +57 -0
  4. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/generation_config.json +10 -0
  5. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/merges.txt +0 -0
  6. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/model.safetensors +3 -0
  7. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/optimizer.pt +3 -0
  8. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/rng_state_0.pth +3 -0
  9. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/rng_state_1.pth +3 -0
  10. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/scheduler.pt +3 -0
  11. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/special_tokens_map.json +37 -0
  12. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/tokenizer.json +0 -0
  13. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/tokenizer_config.json +31 -0
  14. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/trainer_state.json +0 -0
  15. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/training_args.bin +3 -0
  16. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/vocab.json +0 -0
  17. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/added_tokens.json +3 -0
  18. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/chat_template.jinja +11 -0
  19. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/config.json +57 -0
  20. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/generation_config.json +10 -0
  21. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/merges.txt +0 -0
  22. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/model.safetensors +3 -0
  23. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/optimizer.pt +3 -0
  24. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/rng_state_0.pth +3 -0
  25. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/rng_state_1.pth +3 -0
  26. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/scheduler.pt +3 -0
  27. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/special_tokens_map.json +37 -0
  28. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/tokenizer.json +0 -0
  29. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/tokenizer_config.json +31 -0
  30. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/trainer_state.json +0 -0
  31. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/training_args.bin +3 -0
  32. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/vocab.json +0 -0
  33. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/added_tokens.json +3 -0
  34. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/chat_template.jinja +11 -0
  35. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/config.json +57 -0
  36. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/generation_config.json +10 -0
  37. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/merges.txt +0 -0
  38. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/model.safetensors +3 -0
  39. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/optimizer.pt +3 -0
  40. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/rng_state_0.pth +3 -0
  41. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/rng_state_1.pth +3 -0
  42. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/scheduler.pt +3 -0
  43. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/special_tokens_map.json +37 -0
  44. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/tokenizer.json +0 -0
  45. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/tokenizer_config.json +31 -0
  46. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/trainer_state.json +0 -0
  47. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/training_args.bin +3 -0
  48. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/vocab.json +0 -0
  49. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert3/checkpoint-90000/added_tokens.json +3 -0
  50. pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert3/checkpoint-90000/chat_template.jinja +11 -0
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<|mdm_mask|>": 50257
3
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/chat_template.jinja ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {% set loop_messages = messages %}
2
+ {% for message in loop_messages %}
3
+ {% if loop.index0 == 0 %}{{ bos_token }}{% endif %}
4
+ <|start_header_id|>{{ message['role'] }}<|end_header_id|>
5
+
6
+ {{ message['content'] | trim }}<|eot_id|>
7
+ {%- endfor %}
8
+ {% if add_generation_prompt and (loop_messages | length == 0 or loop_messages[-1]['role'] != 'assistant') %}
9
+ <|start_header_id|>assistant<|end_header_id|>
10
+
11
+ {% endif %}
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_type": "silu",
3
+ "alibi": false,
4
+ "alibi_bias_max": 8.0,
5
+ "architectures": [
6
+ "LLaDAModelLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "attention_layer_norm": false,
10
+ "attention_layer_norm_with_affine": true,
11
+ "auto_map": {
12
+ "AutoConfig": "configuration_llada.LLaDAConfig",
13
+ "AutoModel": "modeling_llada.LLaDAModelLM",
14
+ "AutoModelForCausalLM": "modeling_llada.LLaDAModelLM"
15
+ },
16
+ "bias_for_layer_norm": false,
17
+ "block_group_size": 1,
18
+ "block_type": "llama",
19
+ "bos_token_id": 50256,
20
+ "d_model": 1024,
21
+ "dtype": "bfloat16",
22
+ "embedding_dropout": 0.0,
23
+ "embedding_size": 50304,
24
+ "eos_token_id": 50256,
25
+ "flash_attention": false,
26
+ "include_bias": false,
27
+ "include_qkv_bias": false,
28
+ "init_cutoff_factor": null,
29
+ "init_device": "meta",
30
+ "init_fn": "mitchell",
31
+ "init_std": 0.02,
32
+ "input_emb_norm": false,
33
+ "layer_norm_type": "rms",
34
+ "layer_norm_with_affine": true,
35
+ "mask_token_id": 50257,
36
+ "max_sequence_length": 4096,
37
+ "mlp_hidden_size": 4096,
38
+ "mlp_ratio": 4,
39
+ "model_type": "llada",
40
+ "multi_query_attention": null,
41
+ "n_experts": null,
42
+ "n_heads": 16,
43
+ "n_kv_heads": 16,
44
+ "n_layers": 15,
45
+ "pad_token_id": 50256,
46
+ "precision": "amp_bf16",
47
+ "residual_dropout": 0.0,
48
+ "rms_norm_eps": 1e-05,
49
+ "rope": true,
50
+ "rope_full_precision": true,
51
+ "rope_theta": 500000.0,
52
+ "scale_logits": false,
53
+ "transformers_version": "4.57.0",
54
+ "use_cache": false,
55
+ "vocab_size": 50304,
56
+ "weight_tying": false
57
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": [
5
+ 50256
6
+ ],
7
+ "pad_token_id": 50256,
8
+ "transformers_version": "4.57.0",
9
+ "use_cache": false
10
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62fb9c0e670f6186bb9582e368d5934868420b4722caad0e6be8e2951dbaed88
3
+ size 709441248
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a7e51292f0a1c7ce99501645be98a72fa1cae57950480cffaaa2063ed7b5adc
3
+ size 1418967243
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec74f063acf469fe1893b734daeb60fdefe6ea5859a8db039a90be9a4c022c9c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92b9b2b8700e7e248ddcaea008e458a3dfda5ade2c9b67fedb4bb7968a601c1c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e642bffe9916b1dd2a29308e9659ba2eb1b8d41cfbdbfc375c036a98cd9e743
3
+ size 1465
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|endoftext|>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "mask_token": {
17
+ "content": "<|mdm_mask|>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "pad_token": {
24
+ "content": "<|endoftext|>",
25
+ "lstrip": false,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "<|endoftext|>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/tokenizer_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "50256": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "50257": {
13
+ "content": "<|mdm_mask|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ }
20
+ },
21
+ "bos_token": "<|endoftext|>",
22
+ "clean_up_tokenization_spaces": false,
23
+ "eos_token": "<|endoftext|>",
24
+ "extra_special_tokens": {},
25
+ "mask_token": "<|mdm_mask|>",
26
+ "model_max_length": 1024,
27
+ "pad_token": "<|endoftext|>",
28
+ "padding_side": "right",
29
+ "tokenizer_class": "GPT2Tokenizer",
30
+ "unk_token": "<|endoftext|>"
31
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c048223faf850c67bd9ad046fb9d73dd6666e046f9442e6261c5ba14407fac5c
3
+ size 6161
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert0/checkpoint-90000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<|mdm_mask|>": 50257
3
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/chat_template.jinja ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {% set loop_messages = messages %}
2
+ {% for message in loop_messages %}
3
+ {% if loop.index0 == 0 %}{{ bos_token }}{% endif %}
4
+ <|start_header_id|>{{ message['role'] }}<|end_header_id|>
5
+
6
+ {{ message['content'] | trim }}<|eot_id|>
7
+ {%- endfor %}
8
+ {% if add_generation_prompt and (loop_messages | length == 0 or loop_messages[-1]['role'] != 'assistant') %}
9
+ <|start_header_id|>assistant<|end_header_id|>
10
+
11
+ {% endif %}
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_type": "silu",
3
+ "alibi": false,
4
+ "alibi_bias_max": 8.0,
5
+ "architectures": [
6
+ "LLaDAModelLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "attention_layer_norm": false,
10
+ "attention_layer_norm_with_affine": true,
11
+ "auto_map": {
12
+ "AutoConfig": "configuration_llada.LLaDAConfig",
13
+ "AutoModel": "modeling_llada.LLaDAModelLM",
14
+ "AutoModelForCausalLM": "modeling_llada.LLaDAModelLM"
15
+ },
16
+ "bias_for_layer_norm": false,
17
+ "block_group_size": 1,
18
+ "block_type": "llama",
19
+ "bos_token_id": 50256,
20
+ "d_model": 1024,
21
+ "dtype": "bfloat16",
22
+ "embedding_dropout": 0.0,
23
+ "embedding_size": 50304,
24
+ "eos_token_id": 50256,
25
+ "flash_attention": false,
26
+ "include_bias": false,
27
+ "include_qkv_bias": false,
28
+ "init_cutoff_factor": null,
29
+ "init_device": "meta",
30
+ "init_fn": "mitchell",
31
+ "init_std": 0.02,
32
+ "input_emb_norm": false,
33
+ "layer_norm_type": "rms",
34
+ "layer_norm_with_affine": true,
35
+ "mask_token_id": 50257,
36
+ "max_sequence_length": 4096,
37
+ "mlp_hidden_size": 4096,
38
+ "mlp_ratio": 4,
39
+ "model_type": "llada",
40
+ "multi_query_attention": null,
41
+ "n_experts": null,
42
+ "n_heads": 16,
43
+ "n_kv_heads": 16,
44
+ "n_layers": 15,
45
+ "pad_token_id": 50256,
46
+ "precision": "amp_bf16",
47
+ "residual_dropout": 0.0,
48
+ "rms_norm_eps": 1e-05,
49
+ "rope": true,
50
+ "rope_full_precision": true,
51
+ "rope_theta": 500000.0,
52
+ "scale_logits": false,
53
+ "transformers_version": "4.57.0",
54
+ "use_cache": false,
55
+ "vocab_size": 50304,
56
+ "weight_tying": false
57
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": [
5
+ 50256
6
+ ],
7
+ "pad_token_id": 50256,
8
+ "transformers_version": "4.57.0",
9
+ "use_cache": false
10
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d25b68000b776057c94bc15a41bcddf1dd5f05887ad8ac08f8f56c1ffd880c4
3
+ size 709441248
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a953eb0625765d270c018b76282bbf5eb36b41e6f48f1c97394dd5bef265992
3
+ size 1418967243
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec74f063acf469fe1893b734daeb60fdefe6ea5859a8db039a90be9a4c022c9c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92b9b2b8700e7e248ddcaea008e458a3dfda5ade2c9b67fedb4bb7968a601c1c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e642bffe9916b1dd2a29308e9659ba2eb1b8d41cfbdbfc375c036a98cd9e743
3
+ size 1465
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|endoftext|>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "mask_token": {
17
+ "content": "<|mdm_mask|>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "pad_token": {
24
+ "content": "<|endoftext|>",
25
+ "lstrip": false,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "<|endoftext|>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/tokenizer_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "50256": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "50257": {
13
+ "content": "<|mdm_mask|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ }
20
+ },
21
+ "bos_token": "<|endoftext|>",
22
+ "clean_up_tokenization_spaces": false,
23
+ "eos_token": "<|endoftext|>",
24
+ "extra_special_tokens": {},
25
+ "mask_token": "<|mdm_mask|>",
26
+ "model_max_length": 1024,
27
+ "pad_token": "<|endoftext|>",
28
+ "padding_side": "right",
29
+ "tokenizer_class": "GPT2Tokenizer",
30
+ "unk_token": "<|endoftext|>"
31
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6aa57103eafa8502c3654316fe13f69cb1804b974fbe31ccd3ab4c720f5ddac
3
+ size 6161
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert1/checkpoint-90000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<|mdm_mask|>": 50257
3
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/chat_template.jinja ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {% set loop_messages = messages %}
2
+ {% for message in loop_messages %}
3
+ {% if loop.index0 == 0 %}{{ bos_token }}{% endif %}
4
+ <|start_header_id|>{{ message['role'] }}<|end_header_id|>
5
+
6
+ {{ message['content'] | trim }}<|eot_id|>
7
+ {%- endfor %}
8
+ {% if add_generation_prompt and (loop_messages | length == 0 or loop_messages[-1]['role'] != 'assistant') %}
9
+ <|start_header_id|>assistant<|end_header_id|>
10
+
11
+ {% endif %}
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_type": "silu",
3
+ "alibi": false,
4
+ "alibi_bias_max": 8.0,
5
+ "architectures": [
6
+ "LLaDAModelLM"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "attention_layer_norm": false,
10
+ "attention_layer_norm_with_affine": true,
11
+ "auto_map": {
12
+ "AutoConfig": "configuration_llada.LLaDAConfig",
13
+ "AutoModel": "modeling_llada.LLaDAModelLM",
14
+ "AutoModelForCausalLM": "modeling_llada.LLaDAModelLM"
15
+ },
16
+ "bias_for_layer_norm": false,
17
+ "block_group_size": 1,
18
+ "block_type": "llama",
19
+ "bos_token_id": 50256,
20
+ "d_model": 1024,
21
+ "dtype": "bfloat16",
22
+ "embedding_dropout": 0.0,
23
+ "embedding_size": 50304,
24
+ "eos_token_id": 50256,
25
+ "flash_attention": false,
26
+ "include_bias": false,
27
+ "include_qkv_bias": false,
28
+ "init_cutoff_factor": null,
29
+ "init_device": "meta",
30
+ "init_fn": "mitchell",
31
+ "init_std": 0.02,
32
+ "input_emb_norm": false,
33
+ "layer_norm_type": "rms",
34
+ "layer_norm_with_affine": true,
35
+ "mask_token_id": 50257,
36
+ "max_sequence_length": 4096,
37
+ "mlp_hidden_size": 4096,
38
+ "mlp_ratio": 4,
39
+ "model_type": "llada",
40
+ "multi_query_attention": null,
41
+ "n_experts": null,
42
+ "n_heads": 16,
43
+ "n_kv_heads": 16,
44
+ "n_layers": 15,
45
+ "pad_token_id": 50256,
46
+ "precision": "amp_bf16",
47
+ "residual_dropout": 0.0,
48
+ "rms_norm_eps": 1e-05,
49
+ "rope": true,
50
+ "rope_full_precision": true,
51
+ "rope_theta": 500000.0,
52
+ "scale_logits": false,
53
+ "transformers_version": "4.57.0",
54
+ "use_cache": false,
55
+ "vocab_size": 50304,
56
+ "weight_tying": false
57
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": [
5
+ 50256
6
+ ],
7
+ "pad_token_id": 50256,
8
+ "transformers_version": "4.57.0",
9
+ "use_cache": false
10
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eab5dca34e6c115ed96311de61a496d8fdeb82341b6632f5e12a7f3db0b7ea6b
3
+ size 709441248
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dfa0a6fe0b4f4322ab67318b799550e57673dacf7d9cf50c7eb4a2daf62479b2
3
+ size 1418967243
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec74f063acf469fe1893b734daeb60fdefe6ea5859a8db039a90be9a4c022c9c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92b9b2b8700e7e248ddcaea008e458a3dfda5ade2c9b67fedb4bb7968a601c1c
3
+ size 14725
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e642bffe9916b1dd2a29308e9659ba2eb1b8d41cfbdbfc375c036a98cd9e743
3
+ size 1465
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|endoftext|>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "mask_token": {
17
+ "content": "<|mdm_mask|>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "pad_token": {
24
+ "content": "<|endoftext|>",
25
+ "lstrip": false,
26
+ "normalized": true,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "<|endoftext|>",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/tokenizer_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "50256": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "50257": {
13
+ "content": "<|mdm_mask|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ }
20
+ },
21
+ "bos_token": "<|endoftext|>",
22
+ "clean_up_tokenization_spaces": false,
23
+ "eos_token": "<|endoftext|>",
24
+ "extra_special_tokens": {},
25
+ "mask_token": "<|mdm_mask|>",
26
+ "model_max_length": 1024,
27
+ "pad_token": "<|endoftext|>",
28
+ "padding_side": "right",
29
+ "tokenizer_class": "GPT2Tokenizer",
30
+ "unk_token": "<|endoftext|>"
31
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fae512abf10ca02c5cca8a6fe476c79aaa37f4c684a491e18e6f38f4be10fbe
3
+ size 6161
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert2/checkpoint-90000/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert3/checkpoint-90000/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<|mdm_mask|>": 50257
3
+ }
pt-d-llada-gpt2-355M_fineweb-20B-4clusters_24h/expert3/checkpoint-90000/chat_template.jinja ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {% set loop_messages = messages %}
2
+ {% for message in loop_messages %}
3
+ {% if loop.index0 == 0 %}{{ bos_token }}{% endif %}
4
+ <|start_header_id|>{{ message['role'] }}<|end_header_id|>
5
+
6
+ {{ message['content'] | trim }}<|eot_id|>
7
+ {%- endfor %}
8
+ {% if add_generation_prompt and (loop_messages | length == 0 or loop_messages[-1]['role'] != 'assistant') %}
9
+ <|start_header_id|>assistant<|end_header_id|>
10
+
11
+ {% endif %}