Hugging Face's logo

daibeiya
/
ContextLM

Safetensors
Model card Files
xet
Community
daibeiya commited on
Commit
4d46bcb
·
verified ·
1 Parent(s): adddc7c

model upload

Browse files
Files changed (39) hide show
  1. contextlm_gpt2_base/README.md +81 -0
  2. contextlm_gpt2_base/config.json +43 -0
  3. contextlm_gpt2_base/generation_config.json +6 -0
  4. contextlm_gpt2_base/model.safetensors +3 -0
  5. contextlm_gpt2_base/special_tokens_map.json +5 -0
  6. contextlm_gpt2_base/tokenizer.json +0 -0
  7. contextlm_gpt2_base/tokenizer_config.json +20 -0
  8. contextlm_gpt2_base/trainer_state.json +0 -0
  9. contextlm_gpt2_base/vocab.json +0 -0
  10. contextlm_gpt2_large/README.md +81 -0
  11. contextlm_gpt2_large/config.json +43 -0
  12. contextlm_gpt2_large/generation_config.json +6 -0
  13. contextlm_gpt2_large/model.safetensors +3 -0
  14. contextlm_gpt2_large/special_tokens_map.json +5 -0
  15. contextlm_gpt2_large/tokenizer.json +0 -0
  16. contextlm_gpt2_large/tokenizer_config.json +20 -0
  17. contextlm_gpt2_large/trainer_state.json +0 -0
  18. contextlm_gpt2_large/vocab.json +0 -0
  19. contextlm_gpt2_med/README.md +81 -0
  20. contextlm_gpt2_med/config.json +45 -0
  21. contextlm_gpt2_med/generation_config.json +6 -0
  22. contextlm_gpt2_med/model.safetensors +3 -0
  23. contextlm_gpt2_med/special_tokens_map.json +5 -0
  24. contextlm_gpt2_med/tokenizer.json +0 -0
  25. contextlm_gpt2_med/tokenizer_config.json +20 -0
  26. contextlm_gpt2_med/trainer_state.json +0 -0
  27. contextlm_gpt2_med/vocab.json +0 -0
  28. contextlm_gpt2_xl/README.md +81 -0
  29. contextlm_gpt2_xl/config.json +44 -0
  30. contextlm_gpt2_xl/generation_config.json +6 -0
  31. contextlm_gpt2_xl/model-00001-of-00002.safetensors +3 -0
  32. contextlm_gpt2_xl/model-00002-of-00002.safetensors +3 -0
  33. contextlm_gpt2_xl/model.safetensors.index.json +612 -0
  34. contextlm_gpt2_xl/special_tokens_map.json +5 -0
  35. contextlm_gpt2_xl/tokenizer.json +0 -0
  36. contextlm_gpt2_xl/tokenizer_config.json +20 -0
  37. contextlm_gpt2_xl/trainer_state.json +0 -0
  38. contextlm_gpt2_xl/training_args.bin +3 -0
  39. contextlm_gpt2_xl/vocab.json +0 -0
contextlm_gpt2_base/README.md ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ base_model: /fs-computility/plm/linzhouhan/daibeiya/models/gpt2
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - openwebtext
8
+ model-index:
9
+ - name: gpt2_base_contextlm_l0212_add_lnnorm_wodetach_v2_lr_bf16_lr1e-3
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # gpt2_base_contextlm_l0212_add_lnnorm_wodetach_v2_lr_bf16_lr1e-3
17
+
18
+ This model is a fine-tuned version of [/fs-computility/plm/linzhouhan/daibeiya/models/gpt2](https://huggingface.co//fs-computility/plm/linzhouhan/daibeiya/models/gpt2) on the openwebtext dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 3.0300
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.001
40
+ - train_batch_size: 16
41
+ - eval_batch_size: 8
42
+ - seed: 42
43
+ - distributed_type: multi-GPU
44
+ - num_devices: 16
45
+ - gradient_accumulation_steps: 2
46
+ - total_train_batch_size: 512
47
+ - total_eval_batch_size: 128
48
+ - optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_ratio: 0.05
51
+ - num_epochs: 1.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:------:|:-----:|:---------------:|
57
+ | 3.9616 | 0.0580 | 1000 | 3.8911 |
58
+ | 3.5512 | 0.1160 | 2000 | 3.4861 |
59
+ | 3.4279 | 0.1741 | 3000 | 3.3560 |
60
+ | 3.3471 | 0.2321 | 4000 | 3.2811 |
61
+ | 3.2957 | 0.2901 | 5000 | 3.2321 |
62
+ | 3.2677 | 0.3481 | 6000 | 3.1945 |
63
+ | 3.225 | 0.4062 | 7000 | 3.1653 |
64
+ | 3.2051 | 0.4642 | 8000 | 3.1390 |
65
+ | 3.1816 | 0.5222 | 9000 | 3.1161 |
66
+ | 3.1583 | 0.5802 | 10000 | 3.0971 |
67
+ | 3.1464 | 0.6383 | 11000 | 3.0794 |
68
+ | 3.1365 | 0.6963 | 12000 | 3.0645 |
69
+ | 3.1256 | 0.7543 | 13000 | 3.0509 |
70
+ | 3.1073 | 0.8123 | 14000 | 3.0417 |
71
+ | 3.108 | 0.8703 | 15000 | 3.0349 |
72
+ | 3.098 | 0.9284 | 16000 | 3.0312 |
73
+ | 3.092 | 0.9864 | 17000 | 3.0301 |
74
+
75
+
76
+ ### Framework versions
77
+
78
+ - Transformers 4.51.3
79
+ - Pytorch 2.3.0+cu121
80
+ - Datasets 4.0.0
81
+ - Tokenizers 0.21.4
contextlm_gpt2_base/config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "TokenGPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.1,
7
+ "bos_token_id": 50256,
8
+ "context_lm_layers": 2,
9
+ "embd_pdrop": 0.1,
10
+ "encoder_layers": 0,
11
+ "eos_token_id": 50256,
12
+ "hlm_n_embd": 768,
13
+ "hlm_n_head": 12,
14
+ "initializer_range": 0.02,
15
+ "layer_norm_epsilon": 1e-05,
16
+ "loss_type": "",
17
+ "model_type": "gpt2",
18
+ "n_ctx": 1024,
19
+ "n_embd": 768,
20
+ "n_head": 12,
21
+ "n_inner": null,
22
+ "n_layer": 12,
23
+ "n_positions": 1024,
24
+ "reorder_and_upcast_attn": false,
25
+ "resid_pdrop": 0.1,
26
+ "scale_attn_by_inverse_layer_idx": false,
27
+ "scale_attn_weights": true,
28
+ "summary_activation": null,
29
+ "summary_first_dropout": 0.1,
30
+ "summary_proj_to_labels": true,
31
+ "summary_type": "cls_index",
32
+ "summary_use_proj": true,
33
+ "task_specific_params": {
34
+ "text-generation": {
35
+ "do_sample": true,
36
+ "max_length": 50
37
+ }
38
+ },
39
+ "torch_dtype": "float32",
40
+ "use_context_lm": true,
41
+ "vocab_size": 50257,
42
+ "w_size": 4
43
+ }
contextlm_gpt2_base/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.51.3"
6
+ }
contextlm_gpt2_base/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c916b3802836617a1c1adf9fee6a96918451676a6c74c50d71d341fbf8715fd
3
+ size 708869264
contextlm_gpt2_base/special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
contextlm_gpt2_base/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_base/tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ },
13
+ "bos_token": "<|endoftext|>",
14
+ "clean_up_tokenization_spaces": false,
15
+ "eos_token": "<|endoftext|>",
16
+ "extra_special_tokens": {},
17
+ "model_max_length": 1024,
18
+ "tokenizer_class": "GPT2Tokenizer",
19
+ "unk_token": "<|endoftext|>"
20
+ }
contextlm_gpt2_base/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_base/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_large/README.md ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ base_model: /fs-computility/plm/linzhouhan/daibeiya/models/gpt2-large
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - openwebtext
8
+ model-index:
9
+ - name: gpt2_large_contextlm_l0236_add_lnnorm_lr_bf16_lr6e-4
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # gpt2_large_contextlm_l0236_add_lnnorm_lr_bf16_lr6e-4
17
+
18
+ This model is a fine-tuned version of [/fs-computility/plm/linzhouhan/daibeiya/models/gpt2-large](https://huggingface.co//fs-computility/plm/linzhouhan/daibeiya/models/gpt2-large) on the openwebtext dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 2.7357
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.0006
40
+ - train_batch_size: 4
41
+ - eval_batch_size: 8
42
+ - seed: 42
43
+ - distributed_type: multi-GPU
44
+ - num_devices: 64
45
+ - gradient_accumulation_steps: 2
46
+ - total_train_batch_size: 512
47
+ - total_eval_batch_size: 512
48
+ - optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_ratio: 0.05
51
+ - num_epochs: 1.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:------:|:-----:|:---------------:|
57
+ | 3.8376 | 0.0580 | 1000 | 3.7978 |
58
+ | 3.3595 | 0.1160 | 2000 | 3.3179 |
59
+ | 3.207 | 0.1741 | 3000 | 3.1580 |
60
+ | 3.0999 | 0.2321 | 4000 | 3.0631 |
61
+ | 3.0333 | 0.2901 | 5000 | 2.9981 |
62
+ | 2.9958 | 0.3481 | 6000 | 2.9496 |
63
+ | 2.9443 | 0.4062 | 7000 | 2.9102 |
64
+ | 2.9097 | 0.4642 | 8000 | 2.8760 |
65
+ | 2.879 | 0.5222 | 9000 | 2.8451 |
66
+ | 2.8506 | 0.5802 | 10000 | 2.8198 |
67
+ | 2.831 | 0.6383 | 11000 | 2.7969 |
68
+ | 2.8156 | 0.6963 | 12000 | 2.7781 |
69
+ | 2.799 | 0.7543 | 13000 | 2.7616 |
70
+ | 2.7802 | 0.8123 | 14000 | 2.7494 |
71
+ | 2.7785 | 0.8703 | 15000 | 2.7414 |
72
+ | 2.7706 | 0.9284 | 16000 | 2.7370 |
73
+ | 2.7665 | 0.9864 | 17000 | 2.7357 |
74
+
75
+
76
+ ### Framework versions
77
+
78
+ - Transformers 4.51.3
79
+ - Pytorch 2.3.0+cu121
80
+ - Datasets 4.0.0
81
+ - Tokenizers 0.21.4
contextlm_gpt2_large/config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "TokenGPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.1,
7
+ "bos_token_id": 50256,
8
+ "context_lm_layers": 2,
9
+ "embd_pdrop": 0.1,
10
+ "encoder_layers": 0,
11
+ "eos_token_id": 50256,
12
+ "hlm_n_embd": 1280,
13
+ "hlm_n_head": 20,
14
+ "initializer_range": 0.02,
15
+ "layer_norm_epsilon": 1e-05,
16
+ "loss_type": "",
17
+ "model_type": "gpt2",
18
+ "n_ctx": 1024,
19
+ "n_embd": 1280,
20
+ "n_head": 20,
21
+ "n_inner": null,
22
+ "n_layer": 36,
23
+ "n_positions": 1024,
24
+ "reorder_and_upcast_attn": false,
25
+ "resid_pdrop": 0.1,
26
+ "scale_attn_by_inverse_layer_idx": false,
27
+ "scale_attn_weights": true,
28
+ "summary_activation": null,
29
+ "summary_first_dropout": 0.1,
30
+ "summary_proj_to_labels": true,
31
+ "summary_type": "cls_index",
32
+ "summary_use_proj": true,
33
+ "task_specific_params": {
34
+ "text-generation": {
35
+ "do_sample": true,
36
+ "max_length": 50
37
+ }
38
+ },
39
+ "torch_dtype": "float32",
40
+ "use_context_lm": true,
41
+ "vocab_size": 50257,
42
+ "w_size": 4
43
+ }
contextlm_gpt2_large/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.51.3"
6
+ }
contextlm_gpt2_large/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ace6b1c7f74a072624b7e6a60c078b0fd32f7e285fd5740dccfac2215afb2ee4
3
+ size 3510903960
contextlm_gpt2_large/special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
contextlm_gpt2_large/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_large/tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ },
13
+ "bos_token": "<|endoftext|>",
14
+ "clean_up_tokenization_spaces": false,
15
+ "eos_token": "<|endoftext|>",
16
+ "extra_special_tokens": {},
17
+ "model_max_length": 1024,
18
+ "tokenizer_class": "GPT2Tokenizer",
19
+ "unk_token": "<|endoftext|>"
20
+ }
contextlm_gpt2_large/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_large/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_med/README.md ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ base_model: /fs-computility/plm/linzhouhan/daibeiya/models/gpt2-medium
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - openwebtext
8
+ model-index:
9
+ - name: gpt2_med_contextlm_l0224_add_lnsnorm_lr_bf16_lr8e-4
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # gpt2_med_contextlm_l0224_add_lnsnorm_lr_bf16_lr8e-4
17
+
18
+ This model is a fine-tuned version of [/fs-computility/plm/linzhouhan/daibeiya/models/gpt2-medium](https://huggingface.co//fs-computility/plm/linzhouhan/daibeiya/models/gpt2-medium) on the openwebtext dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 2.8359
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.0008
40
+ - train_batch_size: 8
41
+ - eval_batch_size: 8
42
+ - seed: 42
43
+ - distributed_type: multi-GPU
44
+ - num_devices: 32
45
+ - gradient_accumulation_steps: 2
46
+ - total_train_batch_size: 512
47
+ - total_eval_batch_size: 256
48
+ - optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_ratio: 0.05
51
+ - num_epochs: 1.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:------:|:-----:|:---------------:|
57
+ | 3.8765 | 0.0580 | 1000 | 3.8193 |
58
+ | 3.421 | 0.1160 | 2000 | 3.3707 |
59
+ | 3.279 | 0.1741 | 3000 | 3.2197 |
60
+ | 3.1821 | 0.2321 | 4000 | 3.1376 |
61
+ | 3.1191 | 0.2901 | 5000 | 3.0776 |
62
+ | 3.0841 | 0.3481 | 6000 | 3.0317 |
63
+ | 3.0417 | 0.4062 | 7000 | 2.9948 |
64
+ | 3.0114 | 0.4642 | 8000 | 2.9659 |
65
+ | 2.9864 | 0.5222 | 9000 | 2.9386 |
66
+ | 2.9603 | 0.5802 | 10000 | 2.9137 |
67
+ | 2.9408 | 0.6383 | 11000 | 2.8931 |
68
+ | 2.926 | 0.6963 | 12000 | 2.8762 |
69
+ | 2.9121 | 0.7543 | 13000 | 2.8603 |
70
+ | 2.8928 | 0.8123 | 14000 | 2.8493 |
71
+ | 2.8938 | 0.8703 | 15000 | 2.8413 |
72
+ | 2.8893 | 0.9284 | 16000 | 2.8372 |
73
+ | 2.8822 | 0.9864 | 17000 | 2.8359 |
74
+
75
+
76
+ ### Framework versions
77
+
78
+ - Transformers 4.51.3
79
+ - Pytorch 2.3.0+cu121
80
+ - Datasets 4.0.0
81
+ - Tokenizers 0.21.4
contextlm_gpt2_med/config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "TokenGPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.1,
7
+ "bos_token_id": 50256,
8
+ "context_lm_layers": 2,
9
+ "embd_pdrop": 0.1,
10
+ "encoder_layers": 0,
11
+ "eos_token_id": 50256,
12
+ "hlm_n_embd": 1024,
13
+ "hlm_n_head": 16,
14
+ "initializer_range": 0.02,
15
+ "layer_norm_epsilon": 1e-05,
16
+ "loss_type": "",
17
+ "model_type": "gpt2",
18
+ "n_ctx": 1024,
19
+ "n_embd": 1024,
20
+ "n_head": 16,
21
+ "n_inner": null,
22
+ "n_layer": 24,
23
+ "n_positions": 1024,
24
+ "n_special": 0,
25
+ "predict_special_tokens": true,
26
+ "reorder_and_upcast_attn": false,
27
+ "resid_pdrop": 0.1,
28
+ "scale_attn_by_inverse_layer_idx": false,
29
+ "scale_attn_weights": true,
30
+ "summary_activation": null,
31
+ "summary_first_dropout": 0.1,
32
+ "summary_proj_to_labels": true,
33
+ "summary_type": "cls_index",
34
+ "summary_use_proj": true,
35
+ "task_specific_params": {
36
+ "text-generation": {
37
+ "do_sample": true,
38
+ "max_length": 50
39
+ }
40
+ },
41
+ "torch_dtype": "float32",
42
+ "use_context_lm": true,
43
+ "vocab_size": 50257,
44
+ "w_size": 4
45
+ }
contextlm_gpt2_med/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.51.3"
6
+ }
contextlm_gpt2_med/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38870d5befca0c861aaccc44c93c3f73eec904b3193d6d7282e555ed9bdf930b
3
+ size 1725948040
contextlm_gpt2_med/special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
contextlm_gpt2_med/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_med/tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ },
13
+ "bos_token": "<|endoftext|>",
14
+ "clean_up_tokenization_spaces": false,
15
+ "eos_token": "<|endoftext|>",
16
+ "extra_special_tokens": {},
17
+ "model_max_length": 1024,
18
+ "tokenizer_class": "GPT2Tokenizer",
19
+ "unk_token": "<|endoftext|>"
20
+ }
contextlm_gpt2_med/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_med/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_xl/README.md ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ base_model: /fs-computility/plm/linzhouhan/daibeiya/models/gpt2-xl
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - openwebtext
8
+ model-index:
9
+ - name: gpt2_xl_contextlm_l0248_lnsnorm_add_lr_bf16_lr4e-4
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # gpt2_xl_contextlm_l0248_lnsnorm_add_lr_bf16_lr4e-4
17
+
18
+ This model is a fine-tuned version of [/fs-computility/plm/linzhouhan/daibeiya/models/gpt2-xl](https://huggingface.co//fs-computility/plm/linzhouhan/daibeiya/models/gpt2-xl) on the openwebtext dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 2.6815
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.0004
40
+ - train_batch_size: 2
41
+ - eval_batch_size: 8
42
+ - seed: 42
43
+ - distributed_type: multi-GPU
44
+ - num_devices: 64
45
+ - gradient_accumulation_steps: 4
46
+ - total_train_batch_size: 512
47
+ - total_eval_batch_size: 512
48
+ - optimizer: Use adamw_torch with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_ratio: 0.05
51
+ - num_epochs: 1.0
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:------:|:-----:|:---------------:|
57
+ | 3.886 | 0.0580 | 1000 | 3.8423 |
58
+ | 3.3459 | 0.1160 | 2000 | 3.3027 |
59
+ | 3.1809 | 0.1741 | 3000 | 3.1339 |
60
+ | 3.0677 | 0.2321 | 4000 | 3.0360 |
61
+ | 2.9938 | 0.2901 | 5000 | 2.9655 |
62
+ | 2.954 | 0.3481 | 6000 | 2.9108 |
63
+ | 2.8931 | 0.4062 | 7000 | 2.8683 |
64
+ | 2.8584 | 0.4642 | 8000 | 2.8317 |
65
+ | 2.8266 | 0.5222 | 9000 | 2.7985 |
66
+ | 2.7939 | 0.5802 | 10000 | 2.7710 |
67
+ | 2.7718 | 0.6383 | 11000 | 2.7471 |
68
+ | 2.7557 | 0.6963 | 12000 | 2.7263 |
69
+ | 2.7418 | 0.7543 | 13000 | 2.7088 |
70
+ | 2.7167 | 0.8123 | 14000 | 2.6962 |
71
+ | 2.7161 | 0.8704 | 15000 | 2.6872 |
72
+ | 2.7094 | 0.9284 | 16000 | 2.6828 |
73
+ | 2.7068 | 0.9864 | 17000 | 2.6815 |
74
+
75
+
76
+ ### Framework versions
77
+
78
+ - Transformers 4.51.3
79
+ - Pytorch 2.3.0+cu121
80
+ - Datasets 4.0.0
81
+ - Tokenizers 0.21.4
contextlm_gpt2_xl/config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "TokenGPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.1,
7
+ "bos_token_id": 50256,
8
+ "context_lm_layers": 2,
9
+ "embd_pdrop": 0.1,
10
+ "encoder_layers": 0,
11
+ "eos_token_id": 50256,
12
+ "hlm_n_embd": 1600,
13
+ "hlm_n_head": 25,
14
+ "initializer_range": 0.02,
15
+ "layer_norm_epsilon": 1e-05,
16
+ "loss_type": "",
17
+ "model_type": "gpt2",
18
+ "n_ctx": 1024,
19
+ "n_embd": 1600,
20
+ "n_head": 25,
21
+ "n_inner": null,
22
+ "n_layer": 48,
23
+ "n_positions": 1024,
24
+ "output_past": true,
25
+ "reorder_and_upcast_attn": false,
26
+ "resid_pdrop": 0.1,
27
+ "scale_attn_by_inverse_layer_idx": false,
28
+ "scale_attn_weights": true,
29
+ "summary_activation": null,
30
+ "summary_first_dropout": 0.1,
31
+ "summary_proj_to_labels": true,
32
+ "summary_type": "cls_index",
33
+ "summary_use_proj": true,
34
+ "task_specific_params": {
35
+ "text-generation": {
36
+ "do_sample": true,
37
+ "max_length": 50
38
+ }
39
+ },
40
+ "torch_dtype": "float32",
41
+ "use_context_lm": true,
42
+ "vocab_size": 50257,
43
+ "w_size": 4
44
+ }
contextlm_gpt2_xl/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.51.3"
6
+ }
contextlm_gpt2_xl/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:050d8631c04df02c1e144e4fc623797908fb6a2a675d726164d0405e3acf6118
3
+ size 4959894384
contextlm_gpt2_xl/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4dd54c343114d53d15138ef50a72e82f927754f251a04f2af6c92f48b4f622ae
3
+ size 1838185032
contextlm_gpt2_xl/model.safetensors.index.json ADDED
@@ -0,0 +1,612 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 6798016000
4
+ },
5
+ "weight_map": {
6
+ "context_decoder.0.attn.c_attn.bias": "model-00002-of-00002.safetensors",
7
+ "context_decoder.0.attn.c_attn.weight": "model-00002-of-00002.safetensors",
8
+ "context_decoder.0.attn.c_proj.bias": "model-00002-of-00002.safetensors",
9
+ "context_decoder.0.attn.c_proj.weight": "model-00002-of-00002.safetensors",
10
+ "context_decoder.0.ln_1.bias": "model-00002-of-00002.safetensors",
11
+ "context_decoder.0.ln_1.weight": "model-00002-of-00002.safetensors",
12
+ "context_decoder.0.ln_2.bias": "model-00002-of-00002.safetensors",
13
+ "context_decoder.0.ln_2.weight": "model-00002-of-00002.safetensors",
14
+ "context_decoder.0.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
15
+ "context_decoder.0.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
16
+ "context_decoder.0.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
17
+ "context_decoder.0.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
18
+ "context_decoder.1.attn.c_attn.bias": "model-00002-of-00002.safetensors",
19
+ "context_decoder.1.attn.c_attn.weight": "model-00002-of-00002.safetensors",
20
+ "context_decoder.1.attn.c_proj.bias": "model-00002-of-00002.safetensors",
21
+ "context_decoder.1.attn.c_proj.weight": "model-00002-of-00002.safetensors",
22
+ "context_decoder.1.ln_1.bias": "model-00002-of-00002.safetensors",
23
+ "context_decoder.1.ln_1.weight": "model-00002-of-00002.safetensors",
24
+ "context_decoder.1.ln_2.bias": "model-00002-of-00002.safetensors",
25
+ "context_decoder.1.ln_2.weight": "model-00002-of-00002.safetensors",
26
+ "context_decoder.1.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
27
+ "context_decoder.1.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
28
+ "context_decoder.1.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
29
+ "context_decoder.1.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
30
+ "lm_head.weight": "model-00002-of-00002.safetensors",
31
+ "ln_f.bias": "model-00001-of-00002.safetensors",
32
+ "ln_f.weight": "model-00001-of-00002.safetensors",
33
+ "token_decoder.0.attn.c_attn.bias": "model-00001-of-00002.safetensors",
34
+ "token_decoder.0.attn.c_attn.weight": "model-00001-of-00002.safetensors",
35
+ "token_decoder.0.attn.c_proj.bias": "model-00001-of-00002.safetensors",
36
+ "token_decoder.0.attn.c_proj.weight": "model-00001-of-00002.safetensors",
37
+ "token_decoder.0.ln_1.bias": "model-00001-of-00002.safetensors",
38
+ "token_decoder.0.ln_1.weight": "model-00001-of-00002.safetensors",
39
+ "token_decoder.0.ln_2.bias": "model-00001-of-00002.safetensors",
40
+ "token_decoder.0.ln_2.weight": "model-00001-of-00002.safetensors",
41
+ "token_decoder.0.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
42
+ "token_decoder.0.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
43
+ "token_decoder.0.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
44
+ "token_decoder.0.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
45
+ "token_decoder.1.attn.c_attn.bias": "model-00001-of-00002.safetensors",
46
+ "token_decoder.1.attn.c_attn.weight": "model-00001-of-00002.safetensors",
47
+ "token_decoder.1.attn.c_proj.bias": "model-00001-of-00002.safetensors",
48
+ "token_decoder.1.attn.c_proj.weight": "model-00001-of-00002.safetensors",
49
+ "token_decoder.1.ln_1.bias": "model-00001-of-00002.safetensors",
50
+ "token_decoder.1.ln_1.weight": "model-00001-of-00002.safetensors",
51
+ "token_decoder.1.ln_2.bias": "model-00001-of-00002.safetensors",
52
+ "token_decoder.1.ln_2.weight": "model-00001-of-00002.safetensors",
53
+ "token_decoder.1.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
54
+ "token_decoder.1.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
55
+ "token_decoder.1.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
56
+ "token_decoder.1.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
57
+ "token_decoder.10.attn.c_attn.bias": "model-00001-of-00002.safetensors",
58
+ "token_decoder.10.attn.c_attn.weight": "model-00001-of-00002.safetensors",
59
+ "token_decoder.10.attn.c_proj.bias": "model-00001-of-00002.safetensors",
60
+ "token_decoder.10.attn.c_proj.weight": "model-00001-of-00002.safetensors",
61
+ "token_decoder.10.ln_1.bias": "model-00001-of-00002.safetensors",
62
+ "token_decoder.10.ln_1.weight": "model-00001-of-00002.safetensors",
63
+ "token_decoder.10.ln_2.bias": "model-00001-of-00002.safetensors",
64
+ "token_decoder.10.ln_2.weight": "model-00001-of-00002.safetensors",
65
+ "token_decoder.10.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
66
+ "token_decoder.10.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
67
+ "token_decoder.10.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
68
+ "token_decoder.10.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
69
+ "token_decoder.11.attn.c_attn.bias": "model-00001-of-00002.safetensors",
70
+ "token_decoder.11.attn.c_attn.weight": "model-00001-of-00002.safetensors",
71
+ "token_decoder.11.attn.c_proj.bias": "model-00001-of-00002.safetensors",
72
+ "token_decoder.11.attn.c_proj.weight": "model-00001-of-00002.safetensors",
73
+ "token_decoder.11.ln_1.bias": "model-00001-of-00002.safetensors",
74
+ "token_decoder.11.ln_1.weight": "model-00001-of-00002.safetensors",
75
+ "token_decoder.11.ln_2.bias": "model-00001-of-00002.safetensors",
76
+ "token_decoder.11.ln_2.weight": "model-00001-of-00002.safetensors",
77
+ "token_decoder.11.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
78
+ "token_decoder.11.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
79
+ "token_decoder.11.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
80
+ "token_decoder.11.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
81
+ "token_decoder.12.attn.c_attn.bias": "model-00001-of-00002.safetensors",
82
+ "token_decoder.12.attn.c_attn.weight": "model-00001-of-00002.safetensors",
83
+ "token_decoder.12.attn.c_proj.bias": "model-00001-of-00002.safetensors",
84
+ "token_decoder.12.attn.c_proj.weight": "model-00001-of-00002.safetensors",
85
+ "token_decoder.12.ln_1.bias": "model-00001-of-00002.safetensors",
86
+ "token_decoder.12.ln_1.weight": "model-00001-of-00002.safetensors",
87
+ "token_decoder.12.ln_2.bias": "model-00001-of-00002.safetensors",
88
+ "token_decoder.12.ln_2.weight": "model-00001-of-00002.safetensors",
89
+ "token_decoder.12.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
90
+ "token_decoder.12.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
91
+ "token_decoder.12.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
92
+ "token_decoder.12.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
93
+ "token_decoder.13.attn.c_attn.bias": "model-00001-of-00002.safetensors",
94
+ "token_decoder.13.attn.c_attn.weight": "model-00001-of-00002.safetensors",
95
+ "token_decoder.13.attn.c_proj.bias": "model-00001-of-00002.safetensors",
96
+ "token_decoder.13.attn.c_proj.weight": "model-00001-of-00002.safetensors",
97
+ "token_decoder.13.ln_1.bias": "model-00001-of-00002.safetensors",
98
+ "token_decoder.13.ln_1.weight": "model-00001-of-00002.safetensors",
99
+ "token_decoder.13.ln_2.bias": "model-00001-of-00002.safetensors",
100
+ "token_decoder.13.ln_2.weight": "model-00001-of-00002.safetensors",
101
+ "token_decoder.13.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
102
+ "token_decoder.13.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
103
+ "token_decoder.13.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
104
+ "token_decoder.13.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
105
+ "token_decoder.14.attn.c_attn.bias": "model-00001-of-00002.safetensors",
106
+ "token_decoder.14.attn.c_attn.weight": "model-00001-of-00002.safetensors",
107
+ "token_decoder.14.attn.c_proj.bias": "model-00001-of-00002.safetensors",
108
+ "token_decoder.14.attn.c_proj.weight": "model-00001-of-00002.safetensors",
109
+ "token_decoder.14.ln_1.bias": "model-00001-of-00002.safetensors",
110
+ "token_decoder.14.ln_1.weight": "model-00001-of-00002.safetensors",
111
+ "token_decoder.14.ln_2.bias": "model-00001-of-00002.safetensors",
112
+ "token_decoder.14.ln_2.weight": "model-00001-of-00002.safetensors",
113
+ "token_decoder.14.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
114
+ "token_decoder.14.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
115
+ "token_decoder.14.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
116
+ "token_decoder.14.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
117
+ "token_decoder.15.attn.c_attn.bias": "model-00001-of-00002.safetensors",
118
+ "token_decoder.15.attn.c_attn.weight": "model-00001-of-00002.safetensors",
119
+ "token_decoder.15.attn.c_proj.bias": "model-00001-of-00002.safetensors",
120
+ "token_decoder.15.attn.c_proj.weight": "model-00001-of-00002.safetensors",
121
+ "token_decoder.15.ln_1.bias": "model-00001-of-00002.safetensors",
122
+ "token_decoder.15.ln_1.weight": "model-00001-of-00002.safetensors",
123
+ "token_decoder.15.ln_2.bias": "model-00001-of-00002.safetensors",
124
+ "token_decoder.15.ln_2.weight": "model-00001-of-00002.safetensors",
125
+ "token_decoder.15.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
126
+ "token_decoder.15.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
127
+ "token_decoder.15.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
128
+ "token_decoder.15.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
129
+ "token_decoder.16.attn.c_attn.bias": "model-00001-of-00002.safetensors",
130
+ "token_decoder.16.attn.c_attn.weight": "model-00001-of-00002.safetensors",
131
+ "token_decoder.16.attn.c_proj.bias": "model-00001-of-00002.safetensors",
132
+ "token_decoder.16.attn.c_proj.weight": "model-00001-of-00002.safetensors",
133
+ "token_decoder.16.ln_1.bias": "model-00001-of-00002.safetensors",
134
+ "token_decoder.16.ln_1.weight": "model-00001-of-00002.safetensors",
135
+ "token_decoder.16.ln_2.bias": "model-00001-of-00002.safetensors",
136
+ "token_decoder.16.ln_2.weight": "model-00001-of-00002.safetensors",
137
+ "token_decoder.16.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
138
+ "token_decoder.16.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
139
+ "token_decoder.16.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
140
+ "token_decoder.16.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
141
+ "token_decoder.17.attn.c_attn.bias": "model-00001-of-00002.safetensors",
142
+ "token_decoder.17.attn.c_attn.weight": "model-00001-of-00002.safetensors",
143
+ "token_decoder.17.attn.c_proj.bias": "model-00001-of-00002.safetensors",
144
+ "token_decoder.17.attn.c_proj.weight": "model-00001-of-00002.safetensors",
145
+ "token_decoder.17.ln_1.bias": "model-00001-of-00002.safetensors",
146
+ "token_decoder.17.ln_1.weight": "model-00001-of-00002.safetensors",
147
+ "token_decoder.17.ln_2.bias": "model-00001-of-00002.safetensors",
148
+ "token_decoder.17.ln_2.weight": "model-00001-of-00002.safetensors",
149
+ "token_decoder.17.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
150
+ "token_decoder.17.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
151
+ "token_decoder.17.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
152
+ "token_decoder.17.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
153
+ "token_decoder.18.attn.c_attn.bias": "model-00001-of-00002.safetensors",
154
+ "token_decoder.18.attn.c_attn.weight": "model-00001-of-00002.safetensors",
155
+ "token_decoder.18.attn.c_proj.bias": "model-00001-of-00002.safetensors",
156
+ "token_decoder.18.attn.c_proj.weight": "model-00001-of-00002.safetensors",
157
+ "token_decoder.18.ln_1.bias": "model-00001-of-00002.safetensors",
158
+ "token_decoder.18.ln_1.weight": "model-00001-of-00002.safetensors",
159
+ "token_decoder.18.ln_2.bias": "model-00001-of-00002.safetensors",
160
+ "token_decoder.18.ln_2.weight": "model-00001-of-00002.safetensors",
161
+ "token_decoder.18.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
162
+ "token_decoder.18.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
163
+ "token_decoder.18.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
164
+ "token_decoder.18.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
165
+ "token_decoder.19.attn.c_attn.bias": "model-00001-of-00002.safetensors",
166
+ "token_decoder.19.attn.c_attn.weight": "model-00001-of-00002.safetensors",
167
+ "token_decoder.19.attn.c_proj.bias": "model-00001-of-00002.safetensors",
168
+ "token_decoder.19.attn.c_proj.weight": "model-00001-of-00002.safetensors",
169
+ "token_decoder.19.ln_1.bias": "model-00001-of-00002.safetensors",
170
+ "token_decoder.19.ln_1.weight": "model-00001-of-00002.safetensors",
171
+ "token_decoder.19.ln_2.bias": "model-00001-of-00002.safetensors",
172
+ "token_decoder.19.ln_2.weight": "model-00001-of-00002.safetensors",
173
+ "token_decoder.19.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
174
+ "token_decoder.19.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
175
+ "token_decoder.19.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
176
+ "token_decoder.19.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
177
+ "token_decoder.2.attn.c_attn.bias": "model-00001-of-00002.safetensors",
178
+ "token_decoder.2.attn.c_attn.weight": "model-00001-of-00002.safetensors",
179
+ "token_decoder.2.attn.c_proj.bias": "model-00001-of-00002.safetensors",
180
+ "token_decoder.2.attn.c_proj.weight": "model-00001-of-00002.safetensors",
181
+ "token_decoder.2.ln_1.bias": "model-00001-of-00002.safetensors",
182
+ "token_decoder.2.ln_1.weight": "model-00001-of-00002.safetensors",
183
+ "token_decoder.2.ln_2.bias": "model-00001-of-00002.safetensors",
184
+ "token_decoder.2.ln_2.weight": "model-00001-of-00002.safetensors",
185
+ "token_decoder.2.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
186
+ "token_decoder.2.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
187
+ "token_decoder.2.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
188
+ "token_decoder.2.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
189
+ "token_decoder.20.attn.c_attn.bias": "model-00001-of-00002.safetensors",
190
+ "token_decoder.20.attn.c_attn.weight": "model-00001-of-00002.safetensors",
191
+ "token_decoder.20.attn.c_proj.bias": "model-00001-of-00002.safetensors",
192
+ "token_decoder.20.attn.c_proj.weight": "model-00001-of-00002.safetensors",
193
+ "token_decoder.20.ln_1.bias": "model-00001-of-00002.safetensors",
194
+ "token_decoder.20.ln_1.weight": "model-00001-of-00002.safetensors",
195
+ "token_decoder.20.ln_2.bias": "model-00001-of-00002.safetensors",
196
+ "token_decoder.20.ln_2.weight": "model-00001-of-00002.safetensors",
197
+ "token_decoder.20.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
198
+ "token_decoder.20.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
199
+ "token_decoder.20.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
200
+ "token_decoder.20.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
201
+ "token_decoder.21.attn.c_attn.bias": "model-00001-of-00002.safetensors",
202
+ "token_decoder.21.attn.c_attn.weight": "model-00001-of-00002.safetensors",
203
+ "token_decoder.21.attn.c_proj.bias": "model-00001-of-00002.safetensors",
204
+ "token_decoder.21.attn.c_proj.weight": "model-00001-of-00002.safetensors",
205
+ "token_decoder.21.ln_1.bias": "model-00001-of-00002.safetensors",
206
+ "token_decoder.21.ln_1.weight": "model-00001-of-00002.safetensors",
207
+ "token_decoder.21.ln_2.bias": "model-00001-of-00002.safetensors",
208
+ "token_decoder.21.ln_2.weight": "model-00001-of-00002.safetensors",
209
+ "token_decoder.21.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
210
+ "token_decoder.21.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
211
+ "token_decoder.21.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
212
+ "token_decoder.21.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
213
+ "token_decoder.22.attn.c_attn.bias": "model-00001-of-00002.safetensors",
214
+ "token_decoder.22.attn.c_attn.weight": "model-00001-of-00002.safetensors",
215
+ "token_decoder.22.attn.c_proj.bias": "model-00001-of-00002.safetensors",
216
+ "token_decoder.22.attn.c_proj.weight": "model-00001-of-00002.safetensors",
217
+ "token_decoder.22.ln_1.bias": "model-00001-of-00002.safetensors",
218
+ "token_decoder.22.ln_1.weight": "model-00001-of-00002.safetensors",
219
+ "token_decoder.22.ln_2.bias": "model-00001-of-00002.safetensors",
220
+ "token_decoder.22.ln_2.weight": "model-00001-of-00002.safetensors",
221
+ "token_decoder.22.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
222
+ "token_decoder.22.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
223
+ "token_decoder.22.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
224
+ "token_decoder.22.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
225
+ "token_decoder.23.attn.c_attn.bias": "model-00001-of-00002.safetensors",
226
+ "token_decoder.23.attn.c_attn.weight": "model-00001-of-00002.safetensors",
227
+ "token_decoder.23.attn.c_proj.bias": "model-00001-of-00002.safetensors",
228
+ "token_decoder.23.attn.c_proj.weight": "model-00001-of-00002.safetensors",
229
+ "token_decoder.23.ln_1.bias": "model-00001-of-00002.safetensors",
230
+ "token_decoder.23.ln_1.weight": "model-00001-of-00002.safetensors",
231
+ "token_decoder.23.ln_2.bias": "model-00001-of-00002.safetensors",
232
+ "token_decoder.23.ln_2.weight": "model-00001-of-00002.safetensors",
233
+ "token_decoder.23.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
234
+ "token_decoder.23.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
235
+ "token_decoder.23.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
236
+ "token_decoder.23.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
237
+ "token_decoder.24.attn.c_attn.bias": "model-00001-of-00002.safetensors",
238
+ "token_decoder.24.attn.c_attn.weight": "model-00001-of-00002.safetensors",
239
+ "token_decoder.24.attn.c_proj.bias": "model-00001-of-00002.safetensors",
240
+ "token_decoder.24.attn.c_proj.weight": "model-00001-of-00002.safetensors",
241
+ "token_decoder.24.ln_1.bias": "model-00001-of-00002.safetensors",
242
+ "token_decoder.24.ln_1.weight": "model-00001-of-00002.safetensors",
243
+ "token_decoder.24.ln_2.bias": "model-00001-of-00002.safetensors",
244
+ "token_decoder.24.ln_2.weight": "model-00001-of-00002.safetensors",
245
+ "token_decoder.24.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
246
+ "token_decoder.24.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
247
+ "token_decoder.24.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
248
+ "token_decoder.24.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
249
+ "token_decoder.25.attn.c_attn.bias": "model-00001-of-00002.safetensors",
250
+ "token_decoder.25.attn.c_attn.weight": "model-00001-of-00002.safetensors",
251
+ "token_decoder.25.attn.c_proj.bias": "model-00001-of-00002.safetensors",
252
+ "token_decoder.25.attn.c_proj.weight": "model-00001-of-00002.safetensors",
253
+ "token_decoder.25.ln_1.bias": "model-00001-of-00002.safetensors",
254
+ "token_decoder.25.ln_1.weight": "model-00001-of-00002.safetensors",
255
+ "token_decoder.25.ln_2.bias": "model-00001-of-00002.safetensors",
256
+ "token_decoder.25.ln_2.weight": "model-00001-of-00002.safetensors",
257
+ "token_decoder.25.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
258
+ "token_decoder.25.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
259
+ "token_decoder.25.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
260
+ "token_decoder.25.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
261
+ "token_decoder.26.attn.c_attn.bias": "model-00001-of-00002.safetensors",
262
+ "token_decoder.26.attn.c_attn.weight": "model-00001-of-00002.safetensors",
263
+ "token_decoder.26.attn.c_proj.bias": "model-00001-of-00002.safetensors",
264
+ "token_decoder.26.attn.c_proj.weight": "model-00001-of-00002.safetensors",
265
+ "token_decoder.26.ln_1.bias": "model-00001-of-00002.safetensors",
266
+ "token_decoder.26.ln_1.weight": "model-00001-of-00002.safetensors",
267
+ "token_decoder.26.ln_2.bias": "model-00001-of-00002.safetensors",
268
+ "token_decoder.26.ln_2.weight": "model-00001-of-00002.safetensors",
269
+ "token_decoder.26.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
270
+ "token_decoder.26.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
271
+ "token_decoder.26.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
272
+ "token_decoder.26.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
273
+ "token_decoder.27.attn.c_attn.bias": "model-00001-of-00002.safetensors",
274
+ "token_decoder.27.attn.c_attn.weight": "model-00001-of-00002.safetensors",
275
+ "token_decoder.27.attn.c_proj.bias": "model-00001-of-00002.safetensors",
276
+ "token_decoder.27.attn.c_proj.weight": "model-00001-of-00002.safetensors",
277
+ "token_decoder.27.ln_1.bias": "model-00001-of-00002.safetensors",
278
+ "token_decoder.27.ln_1.weight": "model-00001-of-00002.safetensors",
279
+ "token_decoder.27.ln_2.bias": "model-00001-of-00002.safetensors",
280
+ "token_decoder.27.ln_2.weight": "model-00001-of-00002.safetensors",
281
+ "token_decoder.27.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
282
+ "token_decoder.27.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
283
+ "token_decoder.27.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
284
+ "token_decoder.27.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
285
+ "token_decoder.28.attn.c_attn.bias": "model-00001-of-00002.safetensors",
286
+ "token_decoder.28.attn.c_attn.weight": "model-00001-of-00002.safetensors",
287
+ "token_decoder.28.attn.c_proj.bias": "model-00001-of-00002.safetensors",
288
+ "token_decoder.28.attn.c_proj.weight": "model-00001-of-00002.safetensors",
289
+ "token_decoder.28.ln_1.bias": "model-00001-of-00002.safetensors",
290
+ "token_decoder.28.ln_1.weight": "model-00001-of-00002.safetensors",
291
+ "token_decoder.28.ln_2.bias": "model-00001-of-00002.safetensors",
292
+ "token_decoder.28.ln_2.weight": "model-00001-of-00002.safetensors",
293
+ "token_decoder.28.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
294
+ "token_decoder.28.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
295
+ "token_decoder.28.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
296
+ "token_decoder.28.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
297
+ "token_decoder.29.attn.c_attn.bias": "model-00001-of-00002.safetensors",
298
+ "token_decoder.29.attn.c_attn.weight": "model-00001-of-00002.safetensors",
299
+ "token_decoder.29.attn.c_proj.bias": "model-00001-of-00002.safetensors",
300
+ "token_decoder.29.attn.c_proj.weight": "model-00001-of-00002.safetensors",
301
+ "token_decoder.29.ln_1.bias": "model-00001-of-00002.safetensors",
302
+ "token_decoder.29.ln_1.weight": "model-00001-of-00002.safetensors",
303
+ "token_decoder.29.ln_2.bias": "model-00001-of-00002.safetensors",
304
+ "token_decoder.29.ln_2.weight": "model-00001-of-00002.safetensors",
305
+ "token_decoder.29.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
306
+ "token_decoder.29.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
307
+ "token_decoder.29.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
308
+ "token_decoder.29.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
309
+ "token_decoder.3.attn.c_attn.bias": "model-00001-of-00002.safetensors",
310
+ "token_decoder.3.attn.c_attn.weight": "model-00001-of-00002.safetensors",
311
+ "token_decoder.3.attn.c_proj.bias": "model-00001-of-00002.safetensors",
312
+ "token_decoder.3.attn.c_proj.weight": "model-00001-of-00002.safetensors",
313
+ "token_decoder.3.ln_1.bias": "model-00001-of-00002.safetensors",
314
+ "token_decoder.3.ln_1.weight": "model-00001-of-00002.safetensors",
315
+ "token_decoder.3.ln_2.bias": "model-00001-of-00002.safetensors",
316
+ "token_decoder.3.ln_2.weight": "model-00001-of-00002.safetensors",
317
+ "token_decoder.3.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
318
+ "token_decoder.3.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
319
+ "token_decoder.3.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
320
+ "token_decoder.3.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
321
+ "token_decoder.30.attn.c_attn.bias": "model-00001-of-00002.safetensors",
322
+ "token_decoder.30.attn.c_attn.weight": "model-00001-of-00002.safetensors",
323
+ "token_decoder.30.attn.c_proj.bias": "model-00001-of-00002.safetensors",
324
+ "token_decoder.30.attn.c_proj.weight": "model-00001-of-00002.safetensors",
325
+ "token_decoder.30.ln_1.bias": "model-00001-of-00002.safetensors",
326
+ "token_decoder.30.ln_1.weight": "model-00001-of-00002.safetensors",
327
+ "token_decoder.30.ln_2.bias": "model-00001-of-00002.safetensors",
328
+ "token_decoder.30.ln_2.weight": "model-00001-of-00002.safetensors",
329
+ "token_decoder.30.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
330
+ "token_decoder.30.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
331
+ "token_decoder.30.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
332
+ "token_decoder.30.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
333
+ "token_decoder.31.attn.c_attn.bias": "model-00001-of-00002.safetensors",
334
+ "token_decoder.31.attn.c_attn.weight": "model-00001-of-00002.safetensors",
335
+ "token_decoder.31.attn.c_proj.bias": "model-00001-of-00002.safetensors",
336
+ "token_decoder.31.attn.c_proj.weight": "model-00001-of-00002.safetensors",
337
+ "token_decoder.31.ln_1.bias": "model-00001-of-00002.safetensors",
338
+ "token_decoder.31.ln_1.weight": "model-00001-of-00002.safetensors",
339
+ "token_decoder.31.ln_2.bias": "model-00001-of-00002.safetensors",
340
+ "token_decoder.31.ln_2.weight": "model-00001-of-00002.safetensors",
341
+ "token_decoder.31.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
342
+ "token_decoder.31.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
343
+ "token_decoder.31.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
344
+ "token_decoder.31.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
345
+ "token_decoder.32.attn.c_attn.bias": "model-00001-of-00002.safetensors",
346
+ "token_decoder.32.attn.c_attn.weight": "model-00001-of-00002.safetensors",
347
+ "token_decoder.32.attn.c_proj.bias": "model-00001-of-00002.safetensors",
348
+ "token_decoder.32.attn.c_proj.weight": "model-00001-of-00002.safetensors",
349
+ "token_decoder.32.ln_1.bias": "model-00001-of-00002.safetensors",
350
+ "token_decoder.32.ln_1.weight": "model-00001-of-00002.safetensors",
351
+ "token_decoder.32.ln_2.bias": "model-00001-of-00002.safetensors",
352
+ "token_decoder.32.ln_2.weight": "model-00001-of-00002.safetensors",
353
+ "token_decoder.32.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
354
+ "token_decoder.32.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
355
+ "token_decoder.32.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
356
+ "token_decoder.32.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
357
+ "token_decoder.33.attn.c_attn.bias": "model-00001-of-00002.safetensors",
358
+ "token_decoder.33.attn.c_attn.weight": "model-00001-of-00002.safetensors",
359
+ "token_decoder.33.attn.c_proj.bias": "model-00001-of-00002.safetensors",
360
+ "token_decoder.33.attn.c_proj.weight": "model-00001-of-00002.safetensors",
361
+ "token_decoder.33.ln_1.bias": "model-00001-of-00002.safetensors",
362
+ "token_decoder.33.ln_1.weight": "model-00001-of-00002.safetensors",
363
+ "token_decoder.33.ln_2.bias": "model-00001-of-00002.safetensors",
364
+ "token_decoder.33.ln_2.weight": "model-00001-of-00002.safetensors",
365
+ "token_decoder.33.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
366
+ "token_decoder.33.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
367
+ "token_decoder.33.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
368
+ "token_decoder.33.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
369
+ "token_decoder.34.attn.c_attn.bias": "model-00001-of-00002.safetensors",
370
+ "token_decoder.34.attn.c_attn.weight": "model-00001-of-00002.safetensors",
371
+ "token_decoder.34.attn.c_proj.bias": "model-00001-of-00002.safetensors",
372
+ "token_decoder.34.attn.c_proj.weight": "model-00001-of-00002.safetensors",
373
+ "token_decoder.34.ln_1.bias": "model-00001-of-00002.safetensors",
374
+ "token_decoder.34.ln_1.weight": "model-00001-of-00002.safetensors",
375
+ "token_decoder.34.ln_2.bias": "model-00001-of-00002.safetensors",
376
+ "token_decoder.34.ln_2.weight": "model-00001-of-00002.safetensors",
377
+ "token_decoder.34.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
378
+ "token_decoder.34.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
379
+ "token_decoder.34.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
380
+ "token_decoder.34.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
381
+ "token_decoder.35.attn.c_attn.bias": "model-00001-of-00002.safetensors",
382
+ "token_decoder.35.attn.c_attn.weight": "model-00001-of-00002.safetensors",
383
+ "token_decoder.35.attn.c_proj.bias": "model-00001-of-00002.safetensors",
384
+ "token_decoder.35.attn.c_proj.weight": "model-00001-of-00002.safetensors",
385
+ "token_decoder.35.ln_1.bias": "model-00001-of-00002.safetensors",
386
+ "token_decoder.35.ln_1.weight": "model-00001-of-00002.safetensors",
387
+ "token_decoder.35.ln_2.bias": "model-00001-of-00002.safetensors",
388
+ "token_decoder.35.ln_2.weight": "model-00001-of-00002.safetensors",
389
+ "token_decoder.35.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
390
+ "token_decoder.35.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
391
+ "token_decoder.35.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
392
+ "token_decoder.35.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
393
+ "token_decoder.36.attn.c_attn.bias": "model-00001-of-00002.safetensors",
394
+ "token_decoder.36.attn.c_attn.weight": "model-00001-of-00002.safetensors",
395
+ "token_decoder.36.attn.c_proj.bias": "model-00001-of-00002.safetensors",
396
+ "token_decoder.36.attn.c_proj.weight": "model-00001-of-00002.safetensors",
397
+ "token_decoder.36.ln_1.bias": "model-00001-of-00002.safetensors",
398
+ "token_decoder.36.ln_1.weight": "model-00001-of-00002.safetensors",
399
+ "token_decoder.36.ln_2.bias": "model-00001-of-00002.safetensors",
400
+ "token_decoder.36.ln_2.weight": "model-00001-of-00002.safetensors",
401
+ "token_decoder.36.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
402
+ "token_decoder.36.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
403
+ "token_decoder.36.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
404
+ "token_decoder.36.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
405
+ "token_decoder.37.attn.c_attn.bias": "model-00001-of-00002.safetensors",
406
+ "token_decoder.37.attn.c_attn.weight": "model-00001-of-00002.safetensors",
407
+ "token_decoder.37.attn.c_proj.bias": "model-00001-of-00002.safetensors",
408
+ "token_decoder.37.attn.c_proj.weight": "model-00001-of-00002.safetensors",
409
+ "token_decoder.37.ln_1.bias": "model-00001-of-00002.safetensors",
410
+ "token_decoder.37.ln_1.weight": "model-00001-of-00002.safetensors",
411
+ "token_decoder.37.ln_2.bias": "model-00001-of-00002.safetensors",
412
+ "token_decoder.37.ln_2.weight": "model-00001-of-00002.safetensors",
413
+ "token_decoder.37.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
414
+ "token_decoder.37.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
415
+ "token_decoder.37.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
416
+ "token_decoder.37.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
417
+ "token_decoder.38.attn.c_attn.bias": "model-00002-of-00002.safetensors",
418
+ "token_decoder.38.attn.c_attn.weight": "model-00002-of-00002.safetensors",
419
+ "token_decoder.38.attn.c_proj.bias": "model-00002-of-00002.safetensors",
420
+ "token_decoder.38.attn.c_proj.weight": "model-00002-of-00002.safetensors",
421
+ "token_decoder.38.ln_1.bias": "model-00002-of-00002.safetensors",
422
+ "token_decoder.38.ln_1.weight": "model-00002-of-00002.safetensors",
423
+ "token_decoder.38.ln_2.bias": "model-00002-of-00002.safetensors",
424
+ "token_decoder.38.ln_2.weight": "model-00002-of-00002.safetensors",
425
+ "token_decoder.38.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
426
+ "token_decoder.38.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
427
+ "token_decoder.38.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
428
+ "token_decoder.38.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
429
+ "token_decoder.39.attn.c_attn.bias": "model-00002-of-00002.safetensors",
430
+ "token_decoder.39.attn.c_attn.weight": "model-00002-of-00002.safetensors",
431
+ "token_decoder.39.attn.c_proj.bias": "model-00002-of-00002.safetensors",
432
+ "token_decoder.39.attn.c_proj.weight": "model-00002-of-00002.safetensors",
433
+ "token_decoder.39.ln_1.bias": "model-00002-of-00002.safetensors",
434
+ "token_decoder.39.ln_1.weight": "model-00002-of-00002.safetensors",
435
+ "token_decoder.39.ln_2.bias": "model-00002-of-00002.safetensors",
436
+ "token_decoder.39.ln_2.weight": "model-00002-of-00002.safetensors",
437
+ "token_decoder.39.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
438
+ "token_decoder.39.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
439
+ "token_decoder.39.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
440
+ "token_decoder.39.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
441
+ "token_decoder.4.attn.c_attn.bias": "model-00001-of-00002.safetensors",
442
+ "token_decoder.4.attn.c_attn.weight": "model-00001-of-00002.safetensors",
443
+ "token_decoder.4.attn.c_proj.bias": "model-00001-of-00002.safetensors",
444
+ "token_decoder.4.attn.c_proj.weight": "model-00001-of-00002.safetensors",
445
+ "token_decoder.4.ln_1.bias": "model-00001-of-00002.safetensors",
446
+ "token_decoder.4.ln_1.weight": "model-00001-of-00002.safetensors",
447
+ "token_decoder.4.ln_2.bias": "model-00001-of-00002.safetensors",
448
+ "token_decoder.4.ln_2.weight": "model-00001-of-00002.safetensors",
449
+ "token_decoder.4.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
450
+ "token_decoder.4.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
451
+ "token_decoder.4.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
452
+ "token_decoder.4.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
453
+ "token_decoder.40.attn.c_attn.bias": "model-00002-of-00002.safetensors",
454
+ "token_decoder.40.attn.c_attn.weight": "model-00002-of-00002.safetensors",
455
+ "token_decoder.40.attn.c_proj.bias": "model-00002-of-00002.safetensors",
456
+ "token_decoder.40.attn.c_proj.weight": "model-00002-of-00002.safetensors",
457
+ "token_decoder.40.ln_1.bias": "model-00002-of-00002.safetensors",
458
+ "token_decoder.40.ln_1.weight": "model-00002-of-00002.safetensors",
459
+ "token_decoder.40.ln_2.bias": "model-00002-of-00002.safetensors",
460
+ "token_decoder.40.ln_2.weight": "model-00002-of-00002.safetensors",
461
+ "token_decoder.40.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
462
+ "token_decoder.40.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
463
+ "token_decoder.40.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
464
+ "token_decoder.40.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
465
+ "token_decoder.41.attn.c_attn.bias": "model-00002-of-00002.safetensors",
466
+ "token_decoder.41.attn.c_attn.weight": "model-00002-of-00002.safetensors",
467
+ "token_decoder.41.attn.c_proj.bias": "model-00002-of-00002.safetensors",
468
+ "token_decoder.41.attn.c_proj.weight": "model-00002-of-00002.safetensors",
469
+ "token_decoder.41.ln_1.bias": "model-00002-of-00002.safetensors",
470
+ "token_decoder.41.ln_1.weight": "model-00002-of-00002.safetensors",
471
+ "token_decoder.41.ln_2.bias": "model-00002-of-00002.safetensors",
472
+ "token_decoder.41.ln_2.weight": "model-00002-of-00002.safetensors",
473
+ "token_decoder.41.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
474
+ "token_decoder.41.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
475
+ "token_decoder.41.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
476
+ "token_decoder.41.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
477
+ "token_decoder.42.attn.c_attn.bias": "model-00002-of-00002.safetensors",
478
+ "token_decoder.42.attn.c_attn.weight": "model-00002-of-00002.safetensors",
479
+ "token_decoder.42.attn.c_proj.bias": "model-00002-of-00002.safetensors",
480
+ "token_decoder.42.attn.c_proj.weight": "model-00002-of-00002.safetensors",
481
+ "token_decoder.42.ln_1.bias": "model-00002-of-00002.safetensors",
482
+ "token_decoder.42.ln_1.weight": "model-00002-of-00002.safetensors",
483
+ "token_decoder.42.ln_2.bias": "model-00002-of-00002.safetensors",
484
+ "token_decoder.42.ln_2.weight": "model-00002-of-00002.safetensors",
485
+ "token_decoder.42.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
486
+ "token_decoder.42.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
487
+ "token_decoder.42.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
488
+ "token_decoder.42.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
489
+ "token_decoder.43.attn.c_attn.bias": "model-00002-of-00002.safetensors",
490
+ "token_decoder.43.attn.c_attn.weight": "model-00002-of-00002.safetensors",
491
+ "token_decoder.43.attn.c_proj.bias": "model-00002-of-00002.safetensors",
492
+ "token_decoder.43.attn.c_proj.weight": "model-00002-of-00002.safetensors",
493
+ "token_decoder.43.ln_1.bias": "model-00002-of-00002.safetensors",
494
+ "token_decoder.43.ln_1.weight": "model-00002-of-00002.safetensors",
495
+ "token_decoder.43.ln_2.bias": "model-00002-of-00002.safetensors",
496
+ "token_decoder.43.ln_2.weight": "model-00002-of-00002.safetensors",
497
+ "token_decoder.43.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
498
+ "token_decoder.43.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
499
+ "token_decoder.43.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
500
+ "token_decoder.43.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
501
+ "token_decoder.44.attn.c_attn.bias": "model-00002-of-00002.safetensors",
502
+ "token_decoder.44.attn.c_attn.weight": "model-00002-of-00002.safetensors",
503
+ "token_decoder.44.attn.c_proj.bias": "model-00002-of-00002.safetensors",
504
+ "token_decoder.44.attn.c_proj.weight": "model-00002-of-00002.safetensors",
505
+ "token_decoder.44.ln_1.bias": "model-00002-of-00002.safetensors",
506
+ "token_decoder.44.ln_1.weight": "model-00002-of-00002.safetensors",
507
+ "token_decoder.44.ln_2.bias": "model-00002-of-00002.safetensors",
508
+ "token_decoder.44.ln_2.weight": "model-00002-of-00002.safetensors",
509
+ "token_decoder.44.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
510
+ "token_decoder.44.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
511
+ "token_decoder.44.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
512
+ "token_decoder.44.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
513
+ "token_decoder.45.attn.c_attn.bias": "model-00002-of-00002.safetensors",
514
+ "token_decoder.45.attn.c_attn.weight": "model-00002-of-00002.safetensors",
515
+ "token_decoder.45.attn.c_proj.bias": "model-00002-of-00002.safetensors",
516
+ "token_decoder.45.attn.c_proj.weight": "model-00002-of-00002.safetensors",
517
+ "token_decoder.45.ln_1.bias": "model-00002-of-00002.safetensors",
518
+ "token_decoder.45.ln_1.weight": "model-00002-of-00002.safetensors",
519
+ "token_decoder.45.ln_2.bias": "model-00002-of-00002.safetensors",
520
+ "token_decoder.45.ln_2.weight": "model-00002-of-00002.safetensors",
521
+ "token_decoder.45.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
522
+ "token_decoder.45.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
523
+ "token_decoder.45.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
524
+ "token_decoder.45.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
525
+ "token_decoder.46.attn.c_attn.bias": "model-00002-of-00002.safetensors",
526
+ "token_decoder.46.attn.c_attn.weight": "model-00002-of-00002.safetensors",
527
+ "token_decoder.46.attn.c_proj.bias": "model-00002-of-00002.safetensors",
528
+ "token_decoder.46.attn.c_proj.weight": "model-00002-of-00002.safetensors",
529
+ "token_decoder.46.ln_1.bias": "model-00002-of-00002.safetensors",
530
+ "token_decoder.46.ln_1.weight": "model-00002-of-00002.safetensors",
531
+ "token_decoder.46.ln_2.bias": "model-00002-of-00002.safetensors",
532
+ "token_decoder.46.ln_2.weight": "model-00002-of-00002.safetensors",
533
+ "token_decoder.46.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
534
+ "token_decoder.46.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
535
+ "token_decoder.46.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
536
+ "token_decoder.46.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
537
+ "token_decoder.47.attn.c_attn.bias": "model-00002-of-00002.safetensors",
538
+ "token_decoder.47.attn.c_attn.weight": "model-00002-of-00002.safetensors",
539
+ "token_decoder.47.attn.c_proj.bias": "model-00002-of-00002.safetensors",
540
+ "token_decoder.47.attn.c_proj.weight": "model-00002-of-00002.safetensors",
541
+ "token_decoder.47.ln_1.bias": "model-00002-of-00002.safetensors",
542
+ "token_decoder.47.ln_1.weight": "model-00002-of-00002.safetensors",
543
+ "token_decoder.47.ln_2.bias": "model-00002-of-00002.safetensors",
544
+ "token_decoder.47.ln_2.weight": "model-00002-of-00002.safetensors",
545
+ "token_decoder.47.mlp.c_fc.bias": "model-00002-of-00002.safetensors",
546
+ "token_decoder.47.mlp.c_fc.weight": "model-00002-of-00002.safetensors",
547
+ "token_decoder.47.mlp.c_proj.bias": "model-00002-of-00002.safetensors",
548
+ "token_decoder.47.mlp.c_proj.weight": "model-00002-of-00002.safetensors",
549
+ "token_decoder.5.attn.c_attn.bias": "model-00001-of-00002.safetensors",
550
+ "token_decoder.5.attn.c_attn.weight": "model-00001-of-00002.safetensors",
551
+ "token_decoder.5.attn.c_proj.bias": "model-00001-of-00002.safetensors",
552
+ "token_decoder.5.attn.c_proj.weight": "model-00001-of-00002.safetensors",
553
+ "token_decoder.5.ln_1.bias": "model-00001-of-00002.safetensors",
554
+ "token_decoder.5.ln_1.weight": "model-00001-of-00002.safetensors",
555
+ "token_decoder.5.ln_2.bias": "model-00001-of-00002.safetensors",
556
+ "token_decoder.5.ln_2.weight": "model-00001-of-00002.safetensors",
557
+ "token_decoder.5.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
558
+ "token_decoder.5.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
559
+ "token_decoder.5.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
560
+ "token_decoder.5.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
561
+ "token_decoder.6.attn.c_attn.bias": "model-00001-of-00002.safetensors",
562
+ "token_decoder.6.attn.c_attn.weight": "model-00001-of-00002.safetensors",
563
+ "token_decoder.6.attn.c_proj.bias": "model-00001-of-00002.safetensors",
564
+ "token_decoder.6.attn.c_proj.weight": "model-00001-of-00002.safetensors",
565
+ "token_decoder.6.ln_1.bias": "model-00001-of-00002.safetensors",
566
+ "token_decoder.6.ln_1.weight": "model-00001-of-00002.safetensors",
567
+ "token_decoder.6.ln_2.bias": "model-00001-of-00002.safetensors",
568
+ "token_decoder.6.ln_2.weight": "model-00001-of-00002.safetensors",
569
+ "token_decoder.6.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
570
+ "token_decoder.6.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
571
+ "token_decoder.6.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
572
+ "token_decoder.6.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
573
+ "token_decoder.7.attn.c_attn.bias": "model-00001-of-00002.safetensors",
574
+ "token_decoder.7.attn.c_attn.weight": "model-00001-of-00002.safetensors",
575
+ "token_decoder.7.attn.c_proj.bias": "model-00001-of-00002.safetensors",
576
+ "token_decoder.7.attn.c_proj.weight": "model-00001-of-00002.safetensors",
577
+ "token_decoder.7.ln_1.bias": "model-00001-of-00002.safetensors",
578
+ "token_decoder.7.ln_1.weight": "model-00001-of-00002.safetensors",
579
+ "token_decoder.7.ln_2.bias": "model-00001-of-00002.safetensors",
580
+ "token_decoder.7.ln_2.weight": "model-00001-of-00002.safetensors",
581
+ "token_decoder.7.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
582
+ "token_decoder.7.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
583
+ "token_decoder.7.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
584
+ "token_decoder.7.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
585
+ "token_decoder.8.attn.c_attn.bias": "model-00001-of-00002.safetensors",
586
+ "token_decoder.8.attn.c_attn.weight": "model-00001-of-00002.safetensors",
587
+ "token_decoder.8.attn.c_proj.bias": "model-00001-of-00002.safetensors",
588
+ "token_decoder.8.attn.c_proj.weight": "model-00001-of-00002.safetensors",
589
+ "token_decoder.8.ln_1.bias": "model-00001-of-00002.safetensors",
590
+ "token_decoder.8.ln_1.weight": "model-00001-of-00002.safetensors",
591
+ "token_decoder.8.ln_2.bias": "model-00001-of-00002.safetensors",
592
+ "token_decoder.8.ln_2.weight": "model-00001-of-00002.safetensors",
593
+ "token_decoder.8.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
594
+ "token_decoder.8.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
595
+ "token_decoder.8.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
596
+ "token_decoder.8.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
597
+ "token_decoder.9.attn.c_attn.bias": "model-00001-of-00002.safetensors",
598
+ "token_decoder.9.attn.c_attn.weight": "model-00001-of-00002.safetensors",
599
+ "token_decoder.9.attn.c_proj.bias": "model-00001-of-00002.safetensors",
600
+ "token_decoder.9.attn.c_proj.weight": "model-00001-of-00002.safetensors",
601
+ "token_decoder.9.ln_1.bias": "model-00001-of-00002.safetensors",
602
+ "token_decoder.9.ln_1.weight": "model-00001-of-00002.safetensors",
603
+ "token_decoder.9.ln_2.bias": "model-00001-of-00002.safetensors",
604
+ "token_decoder.9.ln_2.weight": "model-00001-of-00002.safetensors",
605
+ "token_decoder.9.mlp.c_fc.bias": "model-00001-of-00002.safetensors",
606
+ "token_decoder.9.mlp.c_fc.weight": "model-00001-of-00002.safetensors",
607
+ "token_decoder.9.mlp.c_proj.bias": "model-00001-of-00002.safetensors",
608
+ "token_decoder.9.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
609
+ "wpe.weight": "model-00001-of-00002.safetensors",
610
+ "wte.weight": "model-00001-of-00002.safetensors"
611
+ }
612
+ }
contextlm_gpt2_xl/special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
contextlm_gpt2_xl/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_xl/tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ },
13
+ "bos_token": "<|endoftext|>",
14
+ "clean_up_tokenization_spaces": false,
15
+ "eos_token": "<|endoftext|>",
16
+ "extra_special_tokens": {},
17
+ "model_max_length": 1024,
18
+ "tokenizer_class": "GPT2Tokenizer",
19
+ "unk_token": "<|endoftext|>"
20
+ }
contextlm_gpt2_xl/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
contextlm_gpt2_xl/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a97880ab37b6430dcc4dd66cf7c42d52550b439fb8816fa28c8d32c8a6722bc
3
+ size 5432
contextlm_gpt2_xl/vocab.json ADDED
The diff for this file is too large to render. See raw diff