rootxhacker commited on
Commit
db28bb4
·
verified ·
1 Parent(s): 63533e8

Upload folder using huggingface_hub

Browse files
Files changed (45) hide show
  1. checkpoint-1000/config.json +30 -0
  2. checkpoint-1000/generation_config.json +7 -0
  3. checkpoint-1000/model.safetensors +3 -0
  4. checkpoint-1000/optimizer.pt +3 -0
  5. checkpoint-1000/rng_state.pth +3 -0
  6. checkpoint-1000/scheduler.pt +3 -0
  7. checkpoint-1000/special_tokens_map.json +24 -0
  8. checkpoint-1000/tokenizer.json +0 -0
  9. checkpoint-1000/tokenizer.model +3 -0
  10. checkpoint-1000/tokenizer_config.json +42 -0
  11. checkpoint-1000/trainer_state.json +934 -0
  12. checkpoint-1000/training_args.bin +3 -0
  13. checkpoint-1390/config.json +30 -0
  14. checkpoint-1390/generation_config.json +7 -0
  15. checkpoint-1390/model.safetensors +3 -0
  16. checkpoint-1390/optimizer.pt +3 -0
  17. checkpoint-1390/rng_state.pth +3 -0
  18. checkpoint-1390/scheduler.pt +3 -0
  19. checkpoint-1390/special_tokens_map.json +24 -0
  20. checkpoint-1390/tokenizer.json +0 -0
  21. checkpoint-1390/tokenizer.model +3 -0
  22. checkpoint-1390/tokenizer_config.json +42 -0
  23. checkpoint-1390/trainer_state.json +1285 -0
  24. checkpoint-1390/training_args.bin +3 -0
  25. checkpoint-500/config.json +30 -0
  26. checkpoint-500/generation_config.json +7 -0
  27. checkpoint-500/model.safetensors +3 -0
  28. checkpoint-500/optimizer.pt +3 -0
  29. checkpoint-500/rng_state.pth +3 -0
  30. checkpoint-500/scheduler.pt +3 -0
  31. checkpoint-500/special_tokens_map.json +24 -0
  32. checkpoint-500/tokenizer.json +0 -0
  33. checkpoint-500/tokenizer.model +3 -0
  34. checkpoint-500/tokenizer_config.json +42 -0
  35. checkpoint-500/trainer_state.json +484 -0
  36. checkpoint-500/training_args.bin +3 -0
  37. config.json +30 -0
  38. generation_config.json +7 -0
  39. model.safetensors +3 -0
  40. runs/Apr25_07-50-36_9c18e885a673/events.out.tfevents.1745567454.9c18e885a673.5280.1 +3 -0
  41. special_tokens_map.json +24 -0
  42. tokenizer.json +0 -0
  43. tokenizer.model +3 -0
  44. tokenizer_config.json +42 -0
  45. training_args.bin +3 -0
checkpoint-1000/config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 1024,
14
+ "max_position_embeddings": 512,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 6,
18
+ "num_hidden_layers": 6,
19
+ "num_key_value_heads": 6,
20
+ "pad_token_id": 0,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-05,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.51.3",
28
+ "use_cache": true,
29
+ "vocab_size": 32000
30
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.51.3"
7
+ }
checkpoint-1000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:beca8853285b236f336fbac03ad15ef15f3068e8d208d412d19cbbb9610615d3
3
+ size 309900448
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ac58bf92c39d95b75004059055dadfa04f7bbb3df354cd096883cf27a74c677
3
+ size 619836730
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2768285b45b2a0c05f6f50bbb8c0287fca6f62a8cde6d1b1f02151ac72ee8dc
3
+ size 14244
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:642460467c212070f1c720f41b86ecc22bd81d3103fe011fd8d45c9133d2edc0
3
+ size 1064
checkpoint-1000/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-1000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1000/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-1000/tokenizer_config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": true,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": true,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "extra_special_tokens": {},
35
+ "legacy": false,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "sp_model_kwargs": {},
39
+ "tokenizer_class": "LlamaTokenizer",
40
+ "unk_token": "<unk>",
41
+ "use_default_system_prompt": false
42
+ }
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,934 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 7.193202146690519,
6
+ "eval_steps": 500,
7
+ "global_step": 1000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.07155635062611806,
14
+ "grad_norm": 3.981292247772217,
15
+ "learning_rate": 1.9870503597122304e-05,
16
+ "loss": 6.5969,
17
+ "mean_token_accuracy": 0.15720859691500663,
18
+ "num_tokens": 249014.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.14311270125223613,
23
+ "grad_norm": 6.078427314758301,
24
+ "learning_rate": 1.9726618705035972e-05,
25
+ "loss": 6.3047,
26
+ "mean_token_accuracy": 0.17161023132503034,
27
+ "num_tokens": 499527.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.2146690518783542,
32
+ "grad_norm": 3.555305242538452,
33
+ "learning_rate": 1.958273381294964e-05,
34
+ "loss": 6.0472,
35
+ "mean_token_accuracy": 0.19869447350502015,
36
+ "num_tokens": 755405.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.28622540250447226,
41
+ "grad_norm": 3.784306764602661,
42
+ "learning_rate": 1.9438848920863313e-05,
43
+ "loss": 5.8397,
44
+ "mean_token_accuracy": 0.22695979550480844,
45
+ "num_tokens": 1011193.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.35778175313059035,
50
+ "grad_norm": 3.2327146530151367,
51
+ "learning_rate": 1.9294964028776978e-05,
52
+ "loss": 5.6695,
53
+ "mean_token_accuracy": 0.2401440929621458,
54
+ "num_tokens": 1258965.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.4293381037567084,
59
+ "grad_norm": 3.420456647872925,
60
+ "learning_rate": 1.915107913669065e-05,
61
+ "loss": 5.4321,
62
+ "mean_token_accuracy": 0.2521901123225689,
63
+ "num_tokens": 1512466.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.5008944543828264,
68
+ "grad_norm": 4.493953227996826,
69
+ "learning_rate": 1.9007194244604318e-05,
70
+ "loss": 5.2609,
71
+ "mean_token_accuracy": 0.26822467111051085,
72
+ "num_tokens": 1760599.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.5724508050089445,
77
+ "grad_norm": 2.4932243824005127,
78
+ "learning_rate": 1.8863309352517986e-05,
79
+ "loss": 5.1328,
80
+ "mean_token_accuracy": 0.275511983782053,
81
+ "num_tokens": 2014670.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.6440071556350626,
86
+ "grad_norm": 2.402729034423828,
87
+ "learning_rate": 1.8719424460431658e-05,
88
+ "loss": 5.0135,
89
+ "mean_token_accuracy": 0.2811047907918692,
90
+ "num_tokens": 2264498.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.7155635062611807,
95
+ "grad_norm": 2.8561453819274902,
96
+ "learning_rate": 1.8575539568345323e-05,
97
+ "loss": 4.8244,
98
+ "mean_token_accuracy": 0.2967699430882931,
99
+ "num_tokens": 2508975.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.7871198568872988,
104
+ "grad_norm": 2.222463846206665,
105
+ "learning_rate": 1.8431654676258995e-05,
106
+ "loss": 4.7404,
107
+ "mean_token_accuracy": 0.30234326496720315,
108
+ "num_tokens": 2760092.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.8586762075134168,
113
+ "grad_norm": 2.0150718688964844,
114
+ "learning_rate": 1.8287769784172664e-05,
115
+ "loss": 4.5925,
116
+ "mean_token_accuracy": 0.31090952083468437,
117
+ "num_tokens": 3011643.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.9302325581395349,
122
+ "grad_norm": 2.499884605407715,
123
+ "learning_rate": 1.8143884892086332e-05,
124
+ "loss": 4.484,
125
+ "mean_token_accuracy": 0.3174323089420795,
126
+ "num_tokens": 3260597.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 1.0071556350626119,
131
+ "grad_norm": 3.81246280670166,
132
+ "learning_rate": 1.8e-05,
133
+ "loss": 4.8272,
134
+ "mean_token_accuracy": 0.32652662367355534,
135
+ "num_tokens": 3521468.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 1.07871198568873,
140
+ "grad_norm": 1.3433386087417603,
141
+ "learning_rate": 1.785611510791367e-05,
142
+ "loss": 4.2527,
143
+ "mean_token_accuracy": 0.3359356351196766,
144
+ "num_tokens": 3764868.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 1.150268336314848,
149
+ "grad_norm": 2.1221790313720703,
150
+ "learning_rate": 1.771223021582734e-05,
151
+ "loss": 4.1608,
152
+ "mean_token_accuracy": 0.33773822858929636,
153
+ "num_tokens": 4013198.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 1.2218246869409661,
158
+ "grad_norm": 1.3530231714248657,
159
+ "learning_rate": 1.756834532374101e-05,
160
+ "loss": 4.0771,
161
+ "mean_token_accuracy": 0.3416416011750698,
162
+ "num_tokens": 4266761.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 1.293381037567084,
167
+ "grad_norm": 8.359506607055664,
168
+ "learning_rate": 1.7424460431654678e-05,
169
+ "loss": 3.9872,
170
+ "mean_token_accuracy": 0.3446764849126339,
171
+ "num_tokens": 4524185.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 1.364937388193202,
176
+ "grad_norm": 0.9995141625404358,
177
+ "learning_rate": 1.7280575539568346e-05,
178
+ "loss": 3.8942,
179
+ "mean_token_accuracy": 0.3535850510001183,
180
+ "num_tokens": 4775262.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 1.4364937388193202,
185
+ "grad_norm": 0.9777733087539673,
186
+ "learning_rate": 1.7136690647482015e-05,
187
+ "loss": 3.8149,
188
+ "mean_token_accuracy": 0.36338763535022733,
189
+ "num_tokens": 5022677.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 1.5080500894454383,
194
+ "grad_norm": 0.9398026466369629,
195
+ "learning_rate": 1.6992805755395686e-05,
196
+ "loss": 3.7274,
197
+ "mean_token_accuracy": 0.37123609259724616,
198
+ "num_tokens": 5271810.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 1.5796064400715564,
203
+ "grad_norm": 3.464630603790283,
204
+ "learning_rate": 1.6848920863309355e-05,
205
+ "loss": 3.7154,
206
+ "mean_token_accuracy": 0.3758062146604061,
207
+ "num_tokens": 5522408.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 1.6511627906976745,
212
+ "grad_norm": 2.104501485824585,
213
+ "learning_rate": 1.6705035971223023e-05,
214
+ "loss": 3.6469,
215
+ "mean_token_accuracy": 0.3832809627056122,
216
+ "num_tokens": 5769231.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 1.7227191413237923,
221
+ "grad_norm": 2.5066070556640625,
222
+ "learning_rate": 1.656115107913669e-05,
223
+ "loss": 3.6434,
224
+ "mean_token_accuracy": 0.38175718635320666,
225
+ "num_tokens": 6020808.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 1.7942754919499104,
230
+ "grad_norm": 0.9507914781570435,
231
+ "learning_rate": 1.641726618705036e-05,
232
+ "loss": 3.5866,
233
+ "mean_token_accuracy": 0.3858716659247875,
234
+ "num_tokens": 6275383.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 1.8658318425760285,
239
+ "grad_norm": 3.4407248497009277,
240
+ "learning_rate": 1.627338129496403e-05,
241
+ "loss": 3.5663,
242
+ "mean_token_accuracy": 0.38949630260467527,
243
+ "num_tokens": 6525182.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 1.9373881932021466,
248
+ "grad_norm": 0.7253276109695435,
249
+ "learning_rate": 1.61294964028777e-05,
250
+ "loss": 3.5233,
251
+ "mean_token_accuracy": 0.391743092238903,
252
+ "num_tokens": 6779756.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 2.0143112701252237,
257
+ "grad_norm": 0.6907745599746704,
258
+ "learning_rate": 1.598561151079137e-05,
259
+ "loss": 3.8364,
260
+ "mean_token_accuracy": 0.395021445140606,
261
+ "num_tokens": 7043995.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 2.085867620751342,
266
+ "grad_norm": 0.9332271218299866,
267
+ "learning_rate": 1.5841726618705037e-05,
268
+ "loss": 3.4772,
269
+ "mean_token_accuracy": 0.3965952657163143,
270
+ "num_tokens": 7297135.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 2.15742397137746,
275
+ "grad_norm": 1.235854983329773,
276
+ "learning_rate": 1.5697841726618706e-05,
277
+ "loss": 3.4414,
278
+ "mean_token_accuracy": 0.3989549733698368,
279
+ "num_tokens": 7545580.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 2.228980322003578,
284
+ "grad_norm": 0.6656942963600159,
285
+ "learning_rate": 1.5553956834532374e-05,
286
+ "loss": 3.4114,
287
+ "mean_token_accuracy": 0.40089740604162216,
288
+ "num_tokens": 7799266.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 2.300536672629696,
293
+ "grad_norm": 0.6185078024864197,
294
+ "learning_rate": 1.5410071942446046e-05,
295
+ "loss": 3.395,
296
+ "mean_token_accuracy": 0.4044420339167118,
297
+ "num_tokens": 8052210.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 2.3720930232558137,
302
+ "grad_norm": 1.1096031665802002,
303
+ "learning_rate": 1.5266187050359714e-05,
304
+ "loss": 3.3804,
305
+ "mean_token_accuracy": 0.40613374635577204,
306
+ "num_tokens": 8307281.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 2.4436493738819323,
311
+ "grad_norm": 0.5653504729270935,
312
+ "learning_rate": 1.5122302158273383e-05,
313
+ "loss": 3.3501,
314
+ "mean_token_accuracy": 0.40836238488554955,
315
+ "num_tokens": 8557428.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 2.51520572450805,
320
+ "grad_norm": 0.5432367920875549,
321
+ "learning_rate": 1.4978417266187051e-05,
322
+ "loss": 3.3467,
323
+ "mean_token_accuracy": 0.40912329107522966,
324
+ "num_tokens": 8808315.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 2.586762075134168,
329
+ "grad_norm": 1.6327078342437744,
330
+ "learning_rate": 1.4834532374100721e-05,
331
+ "loss": 3.3258,
332
+ "mean_token_accuracy": 0.4106706023216248,
333
+ "num_tokens": 9057886.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 2.658318425760286,
338
+ "grad_norm": 0.6323248147964478,
339
+ "learning_rate": 1.469064748201439e-05,
340
+ "loss": 3.298,
341
+ "mean_token_accuracy": 0.41263881102204325,
342
+ "num_tokens": 9309357.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 2.729874776386404,
347
+ "grad_norm": 2.182023525238037,
348
+ "learning_rate": 1.4546762589928058e-05,
349
+ "loss": 3.2819,
350
+ "mean_token_accuracy": 0.4163250602781773,
351
+ "num_tokens": 9558445.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 2.8014311270125223,
356
+ "grad_norm": 0.5517727136611938,
357
+ "learning_rate": 1.4402877697841728e-05,
358
+ "loss": 3.2486,
359
+ "mean_token_accuracy": 0.4189229063689709,
360
+ "num_tokens": 9813513.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 2.8729874776386404,
365
+ "grad_norm": 0.877505362033844,
366
+ "learning_rate": 1.4258992805755397e-05,
367
+ "loss": 3.2428,
368
+ "mean_token_accuracy": 0.4193955764174461,
369
+ "num_tokens": 10059533.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 2.9445438282647585,
374
+ "grad_norm": 0.7058685421943665,
375
+ "learning_rate": 1.4115107913669065e-05,
376
+ "loss": 3.2369,
377
+ "mean_token_accuracy": 0.4203347131609917,
378
+ "num_tokens": 10309900.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 3.0214669051878356,
383
+ "grad_norm": 1.2357401847839355,
384
+ "learning_rate": 1.3971223021582736e-05,
385
+ "loss": 3.563,
386
+ "mean_token_accuracy": 0.4190788792400825,
387
+ "num_tokens": 10568955.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 3.0930232558139537,
392
+ "grad_norm": 0.7918577790260315,
393
+ "learning_rate": 1.3827338129496404e-05,
394
+ "loss": 3.2081,
395
+ "mean_token_accuracy": 0.4243326403200626,
396
+ "num_tokens": 10818378.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 3.1645796064400717,
401
+ "grad_norm": 0.5052043795585632,
402
+ "learning_rate": 1.3683453237410074e-05,
403
+ "loss": 3.2022,
404
+ "mean_token_accuracy": 0.4242256350815296,
405
+ "num_tokens": 11070954.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 3.23613595706619,
410
+ "grad_norm": 0.5182781219482422,
411
+ "learning_rate": 1.353956834532374e-05,
412
+ "loss": 3.199,
413
+ "mean_token_accuracy": 0.4230089582502842,
414
+ "num_tokens": 11323582.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 3.3076923076923075,
419
+ "grad_norm": 1.1913461685180664,
420
+ "learning_rate": 1.3395683453237411e-05,
421
+ "loss": 3.1853,
422
+ "mean_token_accuracy": 0.42622370198369025,
423
+ "num_tokens": 11574310.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 3.3792486583184256,
428
+ "grad_norm": 3.5984549522399902,
429
+ "learning_rate": 1.3251798561151081e-05,
430
+ "loss": 3.1743,
431
+ "mean_token_accuracy": 0.42716887295246125,
432
+ "num_tokens": 11828106.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 3.4508050089445437,
437
+ "grad_norm": 0.8405038118362427,
438
+ "learning_rate": 1.310791366906475e-05,
439
+ "loss": 3.1591,
440
+ "mean_token_accuracy": 0.4280025988817215,
441
+ "num_tokens": 12078899.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 3.5223613595706618,
446
+ "grad_norm": 0.5538778901100159,
447
+ "learning_rate": 1.296402877697842e-05,
448
+ "loss": 3.1318,
449
+ "mean_token_accuracy": 0.43028812259435656,
450
+ "num_tokens": 12331071.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 3.59391771019678,
455
+ "grad_norm": 2.8024959564208984,
456
+ "learning_rate": 1.2820143884892086e-05,
457
+ "loss": 3.1394,
458
+ "mean_token_accuracy": 0.4312668785452843,
459
+ "num_tokens": 12581208.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 3.665474060822898,
464
+ "grad_norm": 0.5277515053749084,
465
+ "learning_rate": 1.2676258992805757e-05,
466
+ "loss": 3.17,
467
+ "mean_token_accuracy": 0.4279266193509102,
468
+ "num_tokens": 12834894.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 3.737030411449016,
473
+ "grad_norm": 0.5472737550735474,
474
+ "learning_rate": 1.2532374100719425e-05,
475
+ "loss": 3.1045,
476
+ "mean_token_accuracy": 0.43450429365038873,
477
+ "num_tokens": 13084018.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 3.808586762075134,
482
+ "grad_norm": 0.5308877229690552,
483
+ "learning_rate": 1.2388489208633095e-05,
484
+ "loss": 3.1107,
485
+ "mean_token_accuracy": 0.43446925580501555,
486
+ "num_tokens": 13329962.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 3.8801431127012522,
491
+ "grad_norm": 0.5113710761070251,
492
+ "learning_rate": 1.2244604316546764e-05,
493
+ "loss": 3.1275,
494
+ "mean_token_accuracy": 0.4315035603940487,
495
+ "num_tokens": 13579569.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 3.9516994633273703,
500
+ "grad_norm": 0.5252534747123718,
501
+ "learning_rate": 1.2100719424460432e-05,
502
+ "loss": 3.0685,
503
+ "mean_token_accuracy": 0.43880158141255377,
504
+ "num_tokens": 13828657.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 4.028622540250447,
509
+ "grad_norm": 1.111735224723816,
510
+ "learning_rate": 1.1956834532374102e-05,
511
+ "loss": 3.4198,
512
+ "mean_token_accuracy": 0.4328605380000138,
513
+ "num_tokens": 14088119.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 4.100178890876565,
518
+ "grad_norm": 1.315172791481018,
519
+ "learning_rate": 1.1812949640287769e-05,
520
+ "loss": 3.1038,
521
+ "mean_token_accuracy": 0.43393718749284743,
522
+ "num_tokens": 14345645.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 4.171735241502684,
527
+ "grad_norm": 5.589102745056152,
528
+ "learning_rate": 1.1669064748201439e-05,
529
+ "loss": 3.0984,
530
+ "mean_token_accuracy": 0.4368622414767742,
531
+ "num_tokens": 14596528.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 4.243291592128801,
536
+ "grad_norm": 0.7761235237121582,
537
+ "learning_rate": 1.152517985611511e-05,
538
+ "loss": 3.0556,
539
+ "mean_token_accuracy": 0.43845293298363686,
540
+ "num_tokens": 14846676.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 4.31484794275492,
545
+ "grad_norm": 0.646345853805542,
546
+ "learning_rate": 1.1381294964028778e-05,
547
+ "loss": 3.0742,
548
+ "mean_token_accuracy": 0.4365877464413643,
549
+ "num_tokens": 15100337.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 4.386404293381037,
554
+ "grad_norm": 0.5311315059661865,
555
+ "learning_rate": 1.1237410071942448e-05,
556
+ "loss": 3.068,
557
+ "mean_token_accuracy": 0.4402126669883728,
558
+ "num_tokens": 15346095.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 4.457960644007156,
563
+ "grad_norm": 0.6322463154792786,
564
+ "learning_rate": 1.1093525179856115e-05,
565
+ "loss": 3.0371,
566
+ "mean_token_accuracy": 0.4426932595670223,
567
+ "num_tokens": 15596167.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 4.529516994633274,
572
+ "grad_norm": 1.5417993068695068,
573
+ "learning_rate": 1.0949640287769785e-05,
574
+ "loss": 3.0351,
575
+ "mean_token_accuracy": 0.44205591902136804,
576
+ "num_tokens": 15846428.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 4.601073345259392,
581
+ "grad_norm": 1.0343070030212402,
582
+ "learning_rate": 1.0805755395683455e-05,
583
+ "loss": 3.0414,
584
+ "mean_token_accuracy": 0.4421232812106609,
585
+ "num_tokens": 16097718.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 4.67262969588551,
590
+ "grad_norm": 1.1514654159545898,
591
+ "learning_rate": 1.0661870503597123e-05,
592
+ "loss": 3.0388,
593
+ "mean_token_accuracy": 0.44131903648376464,
594
+ "num_tokens": 16346673.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 4.7441860465116275,
599
+ "grad_norm": 0.6893991231918335,
600
+ "learning_rate": 1.0517985611510793e-05,
601
+ "loss": 3.0303,
602
+ "mean_token_accuracy": 0.44393994063138964,
603
+ "num_tokens": 16593174.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 4.815742397137746,
608
+ "grad_norm": 0.5472334027290344,
609
+ "learning_rate": 1.037410071942446e-05,
610
+ "loss": 3.0426,
611
+ "mean_token_accuracy": 0.44140187501907346,
612
+ "num_tokens": 16847713.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 4.8872987477638645,
617
+ "grad_norm": 0.5398672223091125,
618
+ "learning_rate": 1.023021582733813e-05,
619
+ "loss": 3.013,
620
+ "mean_token_accuracy": 0.4440676636993885,
621
+ "num_tokens": 17094603.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 4.958855098389982,
626
+ "grad_norm": 0.6706956028938293,
627
+ "learning_rate": 1.00863309352518e-05,
628
+ "loss": 3.0106,
629
+ "mean_token_accuracy": 0.4429888568818569,
630
+ "num_tokens": 17348923.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 5.035778175313059,
635
+ "grad_norm": 0.5257108211517334,
636
+ "learning_rate": 9.942446043165467e-06,
637
+ "loss": 3.3038,
638
+ "mean_token_accuracy": 0.4451724378074088,
639
+ "num_tokens": 17610256.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 5.107334525939177,
644
+ "grad_norm": 1.5951194763183594,
645
+ "learning_rate": 9.798561151079137e-06,
646
+ "loss": 3.0097,
647
+ "mean_token_accuracy": 0.44397315233945844,
648
+ "num_tokens": 17861102.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 5.178890876565295,
653
+ "grad_norm": 0.5681999921798706,
654
+ "learning_rate": 9.654676258992807e-06,
655
+ "loss": 3.0288,
656
+ "mean_token_accuracy": 0.4430039346218109,
657
+ "num_tokens": 18112854.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 5.2504472271914135,
662
+ "grad_norm": 0.506310224533081,
663
+ "learning_rate": 9.510791366906476e-06,
664
+ "loss": 2.9882,
665
+ "mean_token_accuracy": 0.4454295739531517,
666
+ "num_tokens": 18372210.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 5.322003577817531,
671
+ "grad_norm": 1.2095166444778442,
672
+ "learning_rate": 9.366906474820144e-06,
673
+ "loss": 2.9867,
674
+ "mean_token_accuracy": 0.4489475920796394,
675
+ "num_tokens": 18619825.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 5.39355992844365,
680
+ "grad_norm": 2.2138569355010986,
681
+ "learning_rate": 9.223021582733813e-06,
682
+ "loss": 3.0184,
683
+ "mean_token_accuracy": 0.4448120780289173,
684
+ "num_tokens": 18868814.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 5.465116279069767,
689
+ "grad_norm": 0.478185772895813,
690
+ "learning_rate": 9.079136690647483e-06,
691
+ "loss": 2.994,
692
+ "mean_token_accuracy": 0.4470547877252102,
693
+ "num_tokens": 19119481.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 5.536672629695886,
698
+ "grad_norm": 0.5845445990562439,
699
+ "learning_rate": 8.935251798561151e-06,
700
+ "loss": 2.9565,
701
+ "mean_token_accuracy": 0.45172852873802183,
702
+ "num_tokens": 19365354.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 5.608228980322004,
707
+ "grad_norm": 0.5983964800834656,
708
+ "learning_rate": 8.791366906474822e-06,
709
+ "loss": 2.9786,
710
+ "mean_token_accuracy": 0.4476122513413429,
711
+ "num_tokens": 19614440.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 5.679785330948121,
716
+ "grad_norm": 0.515555739402771,
717
+ "learning_rate": 8.64748201438849e-06,
718
+ "loss": 2.9384,
719
+ "mean_token_accuracy": 0.4537165902554989,
720
+ "num_tokens": 19865631.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 5.75134168157424,
725
+ "grad_norm": 0.5007749199867249,
726
+ "learning_rate": 8.503597122302158e-06,
727
+ "loss": 2.9727,
728
+ "mean_token_accuracy": 0.4497027687728405,
729
+ "num_tokens": 20117880.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 5.822898032200357,
734
+ "grad_norm": 0.6489661931991577,
735
+ "learning_rate": 8.359712230215829e-06,
736
+ "loss": 2.9992,
737
+ "mean_token_accuracy": 0.4464651308953762,
738
+ "num_tokens": 20368274.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 5.894454382826476,
743
+ "grad_norm": 0.43792665004730225,
744
+ "learning_rate": 8.215827338129497e-06,
745
+ "loss": 2.9629,
746
+ "mean_token_accuracy": 0.4492500305175781,
747
+ "num_tokens": 20621574.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 5.966010733452594,
752
+ "grad_norm": 4.761158466339111,
753
+ "learning_rate": 8.071942446043165e-06,
754
+ "loss": 2.9565,
755
+ "mean_token_accuracy": 0.45158678814768793,
756
+ "num_tokens": 20872399.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 6.042933810375671,
761
+ "grad_norm": 0.6465599536895752,
762
+ "learning_rate": 7.928057553956836e-06,
763
+ "loss": 3.2616,
764
+ "mean_token_accuracy": 0.44940542665923516,
765
+ "num_tokens": 21128736.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 6.114490161001789,
770
+ "grad_norm": 0.7279312610626221,
771
+ "learning_rate": 7.784172661870504e-06,
772
+ "loss": 2.9458,
773
+ "mean_token_accuracy": 0.4515425696969032,
774
+ "num_tokens": 21379493.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 6.186046511627907,
779
+ "grad_norm": 0.5757459402084351,
780
+ "learning_rate": 7.640287769784174e-06,
781
+ "loss": 2.9526,
782
+ "mean_token_accuracy": 0.45158767998218535,
783
+ "num_tokens": 21630298.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 6.257602862254025,
788
+ "grad_norm": 0.8392621874809265,
789
+ "learning_rate": 7.496402877697843e-06,
790
+ "loss": 2.9665,
791
+ "mean_token_accuracy": 0.44971805810928345,
792
+ "num_tokens": 21880131.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 6.3291592128801435,
797
+ "grad_norm": 1.6503797769546509,
798
+ "learning_rate": 7.352517985611511e-06,
799
+ "loss": 2.951,
800
+ "mean_token_accuracy": 0.4522330954670906,
801
+ "num_tokens": 22131666.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 6.400715563506261,
806
+ "grad_norm": 1.1257638931274414,
807
+ "learning_rate": 7.20863309352518e-06,
808
+ "loss": 2.9598,
809
+ "mean_token_accuracy": 0.4505872465670109,
810
+ "num_tokens": 22385786.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 6.47227191413238,
815
+ "grad_norm": 0.5459461808204651,
816
+ "learning_rate": 7.064748201438849e-06,
817
+ "loss": 2.927,
818
+ "mean_token_accuracy": 0.45370263084769247,
819
+ "num_tokens": 22638350.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 6.543828264758497,
824
+ "grad_norm": 0.5412788987159729,
825
+ "learning_rate": 6.920863309352519e-06,
826
+ "loss": 2.9412,
827
+ "mean_token_accuracy": 0.4531597658991814,
828
+ "num_tokens": 22891362.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 6.615384615384615,
833
+ "grad_norm": 0.7567277550697327,
834
+ "learning_rate": 6.776978417266188e-06,
835
+ "loss": 2.9678,
836
+ "mean_token_accuracy": 0.449733317643404,
837
+ "num_tokens": 23140429.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 6.6869409660107335,
842
+ "grad_norm": 0.5896158814430237,
843
+ "learning_rate": 6.633093525179857e-06,
844
+ "loss": 2.9374,
845
+ "mean_token_accuracy": 0.45310442596673967,
846
+ "num_tokens": 23390979.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 6.758497316636851,
851
+ "grad_norm": 0.6928054690361023,
852
+ "learning_rate": 6.489208633093526e-06,
853
+ "loss": 2.9311,
854
+ "mean_token_accuracy": 0.4540563628077507,
855
+ "num_tokens": 23643194.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 6.83005366726297,
860
+ "grad_norm": 0.5281582474708557,
861
+ "learning_rate": 6.345323741007194e-06,
862
+ "loss": 2.931,
863
+ "mean_token_accuracy": 0.45377019196748736,
864
+ "num_tokens": 23896338.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 6.901610017889087,
869
+ "grad_norm": 0.6179826259613037,
870
+ "learning_rate": 6.2014388489208645e-06,
871
+ "loss": 2.9377,
872
+ "mean_token_accuracy": 0.45448167622089386,
873
+ "num_tokens": 24143728.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 6.973166368515206,
878
+ "grad_norm": 1.0639809370040894,
879
+ "learning_rate": 6.057553956834533e-06,
880
+ "loss": 2.9194,
881
+ "mean_token_accuracy": 0.45466252863407136,
882
+ "num_tokens": 24394294.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 7.0500894454382825,
887
+ "grad_norm": 1.7252315282821655,
888
+ "learning_rate": 5.913669064748202e-06,
889
+ "loss": 3.2321,
890
+ "mean_token_accuracy": 0.4523723830537098,
891
+ "num_tokens": 24654875.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 7.121645796064401,
896
+ "grad_norm": 2.1776368618011475,
897
+ "learning_rate": 5.769784172661871e-06,
898
+ "loss": 2.9234,
899
+ "mean_token_accuracy": 0.4544162482023239,
900
+ "num_tokens": 24908270.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 7.193202146690519,
905
+ "grad_norm": 0.46500054001808167,
906
+ "learning_rate": 5.62589928057554e-06,
907
+ "loss": 2.939,
908
+ "mean_token_accuracy": 0.4552768111228943,
909
+ "num_tokens": 25154564.0,
910
+ "step": 1000
911
+ }
912
+ ],
913
+ "logging_steps": 10,
914
+ "max_steps": 1390,
915
+ "num_input_tokens_seen": 0,
916
+ "num_train_epochs": 10,
917
+ "save_steps": 500,
918
+ "stateful_callbacks": {
919
+ "TrainerControl": {
920
+ "args": {
921
+ "should_epoch_stop": false,
922
+ "should_evaluate": false,
923
+ "should_log": false,
924
+ "should_save": true,
925
+ "should_training_stop": false
926
+ },
927
+ "attributes": {}
928
+ }
929
+ },
930
+ "total_flos": 2.785681407954125e+16,
931
+ "train_batch_size": 64,
932
+ "trial_name": null,
933
+ "trial_params": null
934
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d25b53cc47734bcef31e00edf7734c0385193e3d581c4ba0c2cd909723a664bb
3
+ size 5688
checkpoint-1390/config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 1024,
14
+ "max_position_embeddings": 512,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 6,
18
+ "num_hidden_layers": 6,
19
+ "num_key_value_heads": 6,
20
+ "pad_token_id": 0,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-05,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.51.3",
28
+ "use_cache": true,
29
+ "vocab_size": 32000
30
+ }
checkpoint-1390/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.51.3"
7
+ }
checkpoint-1390/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc1ecc51f6ae0b74b7232e22a94a71d7ffb72ce57a924be13dc83c8af25c6af6
3
+ size 309900448
checkpoint-1390/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef64d0c6f725c8933ea81f99776d641518ebfa4ada283bd4241828b5a11868c0
3
+ size 619836730
checkpoint-1390/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b11a10749bfb1630d95eef94125f4590e8610c579d217d1f158e71ce518d72b
3
+ size 14244
checkpoint-1390/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f971ff7c98fc25babdad4c3408fee59c26d66dacede55416e7972c527396e20
3
+ size 1064
checkpoint-1390/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-1390/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1390/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-1390/tokenizer_config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": true,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": true,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "extra_special_tokens": {},
35
+ "legacy": false,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "sp_model_kwargs": {},
39
+ "tokenizer_class": "LlamaTokenizer",
40
+ "unk_token": "<unk>",
41
+ "use_default_system_prompt": false
42
+ }
checkpoint-1390/trainer_state.json ADDED
@@ -0,0 +1,1285 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 9.994633273703041,
6
+ "eval_steps": 500,
7
+ "global_step": 1390,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.07155635062611806,
14
+ "grad_norm": 3.981292247772217,
15
+ "learning_rate": 1.9870503597122304e-05,
16
+ "loss": 6.5969,
17
+ "mean_token_accuracy": 0.15720859691500663,
18
+ "num_tokens": 249014.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.14311270125223613,
23
+ "grad_norm": 6.078427314758301,
24
+ "learning_rate": 1.9726618705035972e-05,
25
+ "loss": 6.3047,
26
+ "mean_token_accuracy": 0.17161023132503034,
27
+ "num_tokens": 499527.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.2146690518783542,
32
+ "grad_norm": 3.555305242538452,
33
+ "learning_rate": 1.958273381294964e-05,
34
+ "loss": 6.0472,
35
+ "mean_token_accuracy": 0.19869447350502015,
36
+ "num_tokens": 755405.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.28622540250447226,
41
+ "grad_norm": 3.784306764602661,
42
+ "learning_rate": 1.9438848920863313e-05,
43
+ "loss": 5.8397,
44
+ "mean_token_accuracy": 0.22695979550480844,
45
+ "num_tokens": 1011193.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.35778175313059035,
50
+ "grad_norm": 3.2327146530151367,
51
+ "learning_rate": 1.9294964028776978e-05,
52
+ "loss": 5.6695,
53
+ "mean_token_accuracy": 0.2401440929621458,
54
+ "num_tokens": 1258965.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.4293381037567084,
59
+ "grad_norm": 3.420456647872925,
60
+ "learning_rate": 1.915107913669065e-05,
61
+ "loss": 5.4321,
62
+ "mean_token_accuracy": 0.2521901123225689,
63
+ "num_tokens": 1512466.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.5008944543828264,
68
+ "grad_norm": 4.493953227996826,
69
+ "learning_rate": 1.9007194244604318e-05,
70
+ "loss": 5.2609,
71
+ "mean_token_accuracy": 0.26822467111051085,
72
+ "num_tokens": 1760599.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.5724508050089445,
77
+ "grad_norm": 2.4932243824005127,
78
+ "learning_rate": 1.8863309352517986e-05,
79
+ "loss": 5.1328,
80
+ "mean_token_accuracy": 0.275511983782053,
81
+ "num_tokens": 2014670.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.6440071556350626,
86
+ "grad_norm": 2.402729034423828,
87
+ "learning_rate": 1.8719424460431658e-05,
88
+ "loss": 5.0135,
89
+ "mean_token_accuracy": 0.2811047907918692,
90
+ "num_tokens": 2264498.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.7155635062611807,
95
+ "grad_norm": 2.8561453819274902,
96
+ "learning_rate": 1.8575539568345323e-05,
97
+ "loss": 4.8244,
98
+ "mean_token_accuracy": 0.2967699430882931,
99
+ "num_tokens": 2508975.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.7871198568872988,
104
+ "grad_norm": 2.222463846206665,
105
+ "learning_rate": 1.8431654676258995e-05,
106
+ "loss": 4.7404,
107
+ "mean_token_accuracy": 0.30234326496720315,
108
+ "num_tokens": 2760092.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.8586762075134168,
113
+ "grad_norm": 2.0150718688964844,
114
+ "learning_rate": 1.8287769784172664e-05,
115
+ "loss": 4.5925,
116
+ "mean_token_accuracy": 0.31090952083468437,
117
+ "num_tokens": 3011643.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.9302325581395349,
122
+ "grad_norm": 2.499884605407715,
123
+ "learning_rate": 1.8143884892086332e-05,
124
+ "loss": 4.484,
125
+ "mean_token_accuracy": 0.3174323089420795,
126
+ "num_tokens": 3260597.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 1.0071556350626119,
131
+ "grad_norm": 3.81246280670166,
132
+ "learning_rate": 1.8e-05,
133
+ "loss": 4.8272,
134
+ "mean_token_accuracy": 0.32652662367355534,
135
+ "num_tokens": 3521468.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 1.07871198568873,
140
+ "grad_norm": 1.3433386087417603,
141
+ "learning_rate": 1.785611510791367e-05,
142
+ "loss": 4.2527,
143
+ "mean_token_accuracy": 0.3359356351196766,
144
+ "num_tokens": 3764868.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 1.150268336314848,
149
+ "grad_norm": 2.1221790313720703,
150
+ "learning_rate": 1.771223021582734e-05,
151
+ "loss": 4.1608,
152
+ "mean_token_accuracy": 0.33773822858929636,
153
+ "num_tokens": 4013198.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 1.2218246869409661,
158
+ "grad_norm": 1.3530231714248657,
159
+ "learning_rate": 1.756834532374101e-05,
160
+ "loss": 4.0771,
161
+ "mean_token_accuracy": 0.3416416011750698,
162
+ "num_tokens": 4266761.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 1.293381037567084,
167
+ "grad_norm": 8.359506607055664,
168
+ "learning_rate": 1.7424460431654678e-05,
169
+ "loss": 3.9872,
170
+ "mean_token_accuracy": 0.3446764849126339,
171
+ "num_tokens": 4524185.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 1.364937388193202,
176
+ "grad_norm": 0.9995141625404358,
177
+ "learning_rate": 1.7280575539568346e-05,
178
+ "loss": 3.8942,
179
+ "mean_token_accuracy": 0.3535850510001183,
180
+ "num_tokens": 4775262.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 1.4364937388193202,
185
+ "grad_norm": 0.9777733087539673,
186
+ "learning_rate": 1.7136690647482015e-05,
187
+ "loss": 3.8149,
188
+ "mean_token_accuracy": 0.36338763535022733,
189
+ "num_tokens": 5022677.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 1.5080500894454383,
194
+ "grad_norm": 0.9398026466369629,
195
+ "learning_rate": 1.6992805755395686e-05,
196
+ "loss": 3.7274,
197
+ "mean_token_accuracy": 0.37123609259724616,
198
+ "num_tokens": 5271810.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 1.5796064400715564,
203
+ "grad_norm": 3.464630603790283,
204
+ "learning_rate": 1.6848920863309355e-05,
205
+ "loss": 3.7154,
206
+ "mean_token_accuracy": 0.3758062146604061,
207
+ "num_tokens": 5522408.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 1.6511627906976745,
212
+ "grad_norm": 2.104501485824585,
213
+ "learning_rate": 1.6705035971223023e-05,
214
+ "loss": 3.6469,
215
+ "mean_token_accuracy": 0.3832809627056122,
216
+ "num_tokens": 5769231.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 1.7227191413237923,
221
+ "grad_norm": 2.5066070556640625,
222
+ "learning_rate": 1.656115107913669e-05,
223
+ "loss": 3.6434,
224
+ "mean_token_accuracy": 0.38175718635320666,
225
+ "num_tokens": 6020808.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 1.7942754919499104,
230
+ "grad_norm": 0.9507914781570435,
231
+ "learning_rate": 1.641726618705036e-05,
232
+ "loss": 3.5866,
233
+ "mean_token_accuracy": 0.3858716659247875,
234
+ "num_tokens": 6275383.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 1.8658318425760285,
239
+ "grad_norm": 3.4407248497009277,
240
+ "learning_rate": 1.627338129496403e-05,
241
+ "loss": 3.5663,
242
+ "mean_token_accuracy": 0.38949630260467527,
243
+ "num_tokens": 6525182.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 1.9373881932021466,
248
+ "grad_norm": 0.7253276109695435,
249
+ "learning_rate": 1.61294964028777e-05,
250
+ "loss": 3.5233,
251
+ "mean_token_accuracy": 0.391743092238903,
252
+ "num_tokens": 6779756.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 2.0143112701252237,
257
+ "grad_norm": 0.6907745599746704,
258
+ "learning_rate": 1.598561151079137e-05,
259
+ "loss": 3.8364,
260
+ "mean_token_accuracy": 0.395021445140606,
261
+ "num_tokens": 7043995.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 2.085867620751342,
266
+ "grad_norm": 0.9332271218299866,
267
+ "learning_rate": 1.5841726618705037e-05,
268
+ "loss": 3.4772,
269
+ "mean_token_accuracy": 0.3965952657163143,
270
+ "num_tokens": 7297135.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 2.15742397137746,
275
+ "grad_norm": 1.235854983329773,
276
+ "learning_rate": 1.5697841726618706e-05,
277
+ "loss": 3.4414,
278
+ "mean_token_accuracy": 0.3989549733698368,
279
+ "num_tokens": 7545580.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 2.228980322003578,
284
+ "grad_norm": 0.6656942963600159,
285
+ "learning_rate": 1.5553956834532374e-05,
286
+ "loss": 3.4114,
287
+ "mean_token_accuracy": 0.40089740604162216,
288
+ "num_tokens": 7799266.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 2.300536672629696,
293
+ "grad_norm": 0.6185078024864197,
294
+ "learning_rate": 1.5410071942446046e-05,
295
+ "loss": 3.395,
296
+ "mean_token_accuracy": 0.4044420339167118,
297
+ "num_tokens": 8052210.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 2.3720930232558137,
302
+ "grad_norm": 1.1096031665802002,
303
+ "learning_rate": 1.5266187050359714e-05,
304
+ "loss": 3.3804,
305
+ "mean_token_accuracy": 0.40613374635577204,
306
+ "num_tokens": 8307281.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 2.4436493738819323,
311
+ "grad_norm": 0.5653504729270935,
312
+ "learning_rate": 1.5122302158273383e-05,
313
+ "loss": 3.3501,
314
+ "mean_token_accuracy": 0.40836238488554955,
315
+ "num_tokens": 8557428.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 2.51520572450805,
320
+ "grad_norm": 0.5432367920875549,
321
+ "learning_rate": 1.4978417266187051e-05,
322
+ "loss": 3.3467,
323
+ "mean_token_accuracy": 0.40912329107522966,
324
+ "num_tokens": 8808315.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 2.586762075134168,
329
+ "grad_norm": 1.6327078342437744,
330
+ "learning_rate": 1.4834532374100721e-05,
331
+ "loss": 3.3258,
332
+ "mean_token_accuracy": 0.4106706023216248,
333
+ "num_tokens": 9057886.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 2.658318425760286,
338
+ "grad_norm": 0.6323248147964478,
339
+ "learning_rate": 1.469064748201439e-05,
340
+ "loss": 3.298,
341
+ "mean_token_accuracy": 0.41263881102204325,
342
+ "num_tokens": 9309357.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 2.729874776386404,
347
+ "grad_norm": 2.182023525238037,
348
+ "learning_rate": 1.4546762589928058e-05,
349
+ "loss": 3.2819,
350
+ "mean_token_accuracy": 0.4163250602781773,
351
+ "num_tokens": 9558445.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 2.8014311270125223,
356
+ "grad_norm": 0.5517727136611938,
357
+ "learning_rate": 1.4402877697841728e-05,
358
+ "loss": 3.2486,
359
+ "mean_token_accuracy": 0.4189229063689709,
360
+ "num_tokens": 9813513.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 2.8729874776386404,
365
+ "grad_norm": 0.877505362033844,
366
+ "learning_rate": 1.4258992805755397e-05,
367
+ "loss": 3.2428,
368
+ "mean_token_accuracy": 0.4193955764174461,
369
+ "num_tokens": 10059533.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 2.9445438282647585,
374
+ "grad_norm": 0.7058685421943665,
375
+ "learning_rate": 1.4115107913669065e-05,
376
+ "loss": 3.2369,
377
+ "mean_token_accuracy": 0.4203347131609917,
378
+ "num_tokens": 10309900.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 3.0214669051878356,
383
+ "grad_norm": 1.2357401847839355,
384
+ "learning_rate": 1.3971223021582736e-05,
385
+ "loss": 3.563,
386
+ "mean_token_accuracy": 0.4190788792400825,
387
+ "num_tokens": 10568955.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 3.0930232558139537,
392
+ "grad_norm": 0.7918577790260315,
393
+ "learning_rate": 1.3827338129496404e-05,
394
+ "loss": 3.2081,
395
+ "mean_token_accuracy": 0.4243326403200626,
396
+ "num_tokens": 10818378.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 3.1645796064400717,
401
+ "grad_norm": 0.5052043795585632,
402
+ "learning_rate": 1.3683453237410074e-05,
403
+ "loss": 3.2022,
404
+ "mean_token_accuracy": 0.4242256350815296,
405
+ "num_tokens": 11070954.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 3.23613595706619,
410
+ "grad_norm": 0.5182781219482422,
411
+ "learning_rate": 1.353956834532374e-05,
412
+ "loss": 3.199,
413
+ "mean_token_accuracy": 0.4230089582502842,
414
+ "num_tokens": 11323582.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 3.3076923076923075,
419
+ "grad_norm": 1.1913461685180664,
420
+ "learning_rate": 1.3395683453237411e-05,
421
+ "loss": 3.1853,
422
+ "mean_token_accuracy": 0.42622370198369025,
423
+ "num_tokens": 11574310.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 3.3792486583184256,
428
+ "grad_norm": 3.5984549522399902,
429
+ "learning_rate": 1.3251798561151081e-05,
430
+ "loss": 3.1743,
431
+ "mean_token_accuracy": 0.42716887295246125,
432
+ "num_tokens": 11828106.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 3.4508050089445437,
437
+ "grad_norm": 0.8405038118362427,
438
+ "learning_rate": 1.310791366906475e-05,
439
+ "loss": 3.1591,
440
+ "mean_token_accuracy": 0.4280025988817215,
441
+ "num_tokens": 12078899.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 3.5223613595706618,
446
+ "grad_norm": 0.5538778901100159,
447
+ "learning_rate": 1.296402877697842e-05,
448
+ "loss": 3.1318,
449
+ "mean_token_accuracy": 0.43028812259435656,
450
+ "num_tokens": 12331071.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 3.59391771019678,
455
+ "grad_norm": 2.8024959564208984,
456
+ "learning_rate": 1.2820143884892086e-05,
457
+ "loss": 3.1394,
458
+ "mean_token_accuracy": 0.4312668785452843,
459
+ "num_tokens": 12581208.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 3.665474060822898,
464
+ "grad_norm": 0.5277515053749084,
465
+ "learning_rate": 1.2676258992805757e-05,
466
+ "loss": 3.17,
467
+ "mean_token_accuracy": 0.4279266193509102,
468
+ "num_tokens": 12834894.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 3.737030411449016,
473
+ "grad_norm": 0.5472737550735474,
474
+ "learning_rate": 1.2532374100719425e-05,
475
+ "loss": 3.1045,
476
+ "mean_token_accuracy": 0.43450429365038873,
477
+ "num_tokens": 13084018.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 3.808586762075134,
482
+ "grad_norm": 0.5308877229690552,
483
+ "learning_rate": 1.2388489208633095e-05,
484
+ "loss": 3.1107,
485
+ "mean_token_accuracy": 0.43446925580501555,
486
+ "num_tokens": 13329962.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 3.8801431127012522,
491
+ "grad_norm": 0.5113710761070251,
492
+ "learning_rate": 1.2244604316546764e-05,
493
+ "loss": 3.1275,
494
+ "mean_token_accuracy": 0.4315035603940487,
495
+ "num_tokens": 13579569.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 3.9516994633273703,
500
+ "grad_norm": 0.5252534747123718,
501
+ "learning_rate": 1.2100719424460432e-05,
502
+ "loss": 3.0685,
503
+ "mean_token_accuracy": 0.43880158141255377,
504
+ "num_tokens": 13828657.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 4.028622540250447,
509
+ "grad_norm": 1.111735224723816,
510
+ "learning_rate": 1.1956834532374102e-05,
511
+ "loss": 3.4198,
512
+ "mean_token_accuracy": 0.4328605380000138,
513
+ "num_tokens": 14088119.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 4.100178890876565,
518
+ "grad_norm": 1.315172791481018,
519
+ "learning_rate": 1.1812949640287769e-05,
520
+ "loss": 3.1038,
521
+ "mean_token_accuracy": 0.43393718749284743,
522
+ "num_tokens": 14345645.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 4.171735241502684,
527
+ "grad_norm": 5.589102745056152,
528
+ "learning_rate": 1.1669064748201439e-05,
529
+ "loss": 3.0984,
530
+ "mean_token_accuracy": 0.4368622414767742,
531
+ "num_tokens": 14596528.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 4.243291592128801,
536
+ "grad_norm": 0.7761235237121582,
537
+ "learning_rate": 1.152517985611511e-05,
538
+ "loss": 3.0556,
539
+ "mean_token_accuracy": 0.43845293298363686,
540
+ "num_tokens": 14846676.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 4.31484794275492,
545
+ "grad_norm": 0.646345853805542,
546
+ "learning_rate": 1.1381294964028778e-05,
547
+ "loss": 3.0742,
548
+ "mean_token_accuracy": 0.4365877464413643,
549
+ "num_tokens": 15100337.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 4.386404293381037,
554
+ "grad_norm": 0.5311315059661865,
555
+ "learning_rate": 1.1237410071942448e-05,
556
+ "loss": 3.068,
557
+ "mean_token_accuracy": 0.4402126669883728,
558
+ "num_tokens": 15346095.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 4.457960644007156,
563
+ "grad_norm": 0.6322463154792786,
564
+ "learning_rate": 1.1093525179856115e-05,
565
+ "loss": 3.0371,
566
+ "mean_token_accuracy": 0.4426932595670223,
567
+ "num_tokens": 15596167.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 4.529516994633274,
572
+ "grad_norm": 1.5417993068695068,
573
+ "learning_rate": 1.0949640287769785e-05,
574
+ "loss": 3.0351,
575
+ "mean_token_accuracy": 0.44205591902136804,
576
+ "num_tokens": 15846428.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 4.601073345259392,
581
+ "grad_norm": 1.0343070030212402,
582
+ "learning_rate": 1.0805755395683455e-05,
583
+ "loss": 3.0414,
584
+ "mean_token_accuracy": 0.4421232812106609,
585
+ "num_tokens": 16097718.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 4.67262969588551,
590
+ "grad_norm": 1.1514654159545898,
591
+ "learning_rate": 1.0661870503597123e-05,
592
+ "loss": 3.0388,
593
+ "mean_token_accuracy": 0.44131903648376464,
594
+ "num_tokens": 16346673.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 4.7441860465116275,
599
+ "grad_norm": 0.6893991231918335,
600
+ "learning_rate": 1.0517985611510793e-05,
601
+ "loss": 3.0303,
602
+ "mean_token_accuracy": 0.44393994063138964,
603
+ "num_tokens": 16593174.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 4.815742397137746,
608
+ "grad_norm": 0.5472334027290344,
609
+ "learning_rate": 1.037410071942446e-05,
610
+ "loss": 3.0426,
611
+ "mean_token_accuracy": 0.44140187501907346,
612
+ "num_tokens": 16847713.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 4.8872987477638645,
617
+ "grad_norm": 0.5398672223091125,
618
+ "learning_rate": 1.023021582733813e-05,
619
+ "loss": 3.013,
620
+ "mean_token_accuracy": 0.4440676636993885,
621
+ "num_tokens": 17094603.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 4.958855098389982,
626
+ "grad_norm": 0.6706956028938293,
627
+ "learning_rate": 1.00863309352518e-05,
628
+ "loss": 3.0106,
629
+ "mean_token_accuracy": 0.4429888568818569,
630
+ "num_tokens": 17348923.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 5.035778175313059,
635
+ "grad_norm": 0.5257108211517334,
636
+ "learning_rate": 9.942446043165467e-06,
637
+ "loss": 3.3038,
638
+ "mean_token_accuracy": 0.4451724378074088,
639
+ "num_tokens": 17610256.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 5.107334525939177,
644
+ "grad_norm": 1.5951194763183594,
645
+ "learning_rate": 9.798561151079137e-06,
646
+ "loss": 3.0097,
647
+ "mean_token_accuracy": 0.44397315233945844,
648
+ "num_tokens": 17861102.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 5.178890876565295,
653
+ "grad_norm": 0.5681999921798706,
654
+ "learning_rate": 9.654676258992807e-06,
655
+ "loss": 3.0288,
656
+ "mean_token_accuracy": 0.4430039346218109,
657
+ "num_tokens": 18112854.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 5.2504472271914135,
662
+ "grad_norm": 0.506310224533081,
663
+ "learning_rate": 9.510791366906476e-06,
664
+ "loss": 2.9882,
665
+ "mean_token_accuracy": 0.4454295739531517,
666
+ "num_tokens": 18372210.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 5.322003577817531,
671
+ "grad_norm": 1.2095166444778442,
672
+ "learning_rate": 9.366906474820144e-06,
673
+ "loss": 2.9867,
674
+ "mean_token_accuracy": 0.4489475920796394,
675
+ "num_tokens": 18619825.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 5.39355992844365,
680
+ "grad_norm": 2.2138569355010986,
681
+ "learning_rate": 9.223021582733813e-06,
682
+ "loss": 3.0184,
683
+ "mean_token_accuracy": 0.4448120780289173,
684
+ "num_tokens": 18868814.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 5.465116279069767,
689
+ "grad_norm": 0.478185772895813,
690
+ "learning_rate": 9.079136690647483e-06,
691
+ "loss": 2.994,
692
+ "mean_token_accuracy": 0.4470547877252102,
693
+ "num_tokens": 19119481.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 5.536672629695886,
698
+ "grad_norm": 0.5845445990562439,
699
+ "learning_rate": 8.935251798561151e-06,
700
+ "loss": 2.9565,
701
+ "mean_token_accuracy": 0.45172852873802183,
702
+ "num_tokens": 19365354.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 5.608228980322004,
707
+ "grad_norm": 0.5983964800834656,
708
+ "learning_rate": 8.791366906474822e-06,
709
+ "loss": 2.9786,
710
+ "mean_token_accuracy": 0.4476122513413429,
711
+ "num_tokens": 19614440.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 5.679785330948121,
716
+ "grad_norm": 0.515555739402771,
717
+ "learning_rate": 8.64748201438849e-06,
718
+ "loss": 2.9384,
719
+ "mean_token_accuracy": 0.4537165902554989,
720
+ "num_tokens": 19865631.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 5.75134168157424,
725
+ "grad_norm": 0.5007749199867249,
726
+ "learning_rate": 8.503597122302158e-06,
727
+ "loss": 2.9727,
728
+ "mean_token_accuracy": 0.4497027687728405,
729
+ "num_tokens": 20117880.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 5.822898032200357,
734
+ "grad_norm": 0.6489661931991577,
735
+ "learning_rate": 8.359712230215829e-06,
736
+ "loss": 2.9992,
737
+ "mean_token_accuracy": 0.4464651308953762,
738
+ "num_tokens": 20368274.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 5.894454382826476,
743
+ "grad_norm": 0.43792665004730225,
744
+ "learning_rate": 8.215827338129497e-06,
745
+ "loss": 2.9629,
746
+ "mean_token_accuracy": 0.4492500305175781,
747
+ "num_tokens": 20621574.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 5.966010733452594,
752
+ "grad_norm": 4.761158466339111,
753
+ "learning_rate": 8.071942446043165e-06,
754
+ "loss": 2.9565,
755
+ "mean_token_accuracy": 0.45158678814768793,
756
+ "num_tokens": 20872399.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 6.042933810375671,
761
+ "grad_norm": 0.6465599536895752,
762
+ "learning_rate": 7.928057553956836e-06,
763
+ "loss": 3.2616,
764
+ "mean_token_accuracy": 0.44940542665923516,
765
+ "num_tokens": 21128736.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 6.114490161001789,
770
+ "grad_norm": 0.7279312610626221,
771
+ "learning_rate": 7.784172661870504e-06,
772
+ "loss": 2.9458,
773
+ "mean_token_accuracy": 0.4515425696969032,
774
+ "num_tokens": 21379493.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 6.186046511627907,
779
+ "grad_norm": 0.5757459402084351,
780
+ "learning_rate": 7.640287769784174e-06,
781
+ "loss": 2.9526,
782
+ "mean_token_accuracy": 0.45158767998218535,
783
+ "num_tokens": 21630298.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 6.257602862254025,
788
+ "grad_norm": 0.8392621874809265,
789
+ "learning_rate": 7.496402877697843e-06,
790
+ "loss": 2.9665,
791
+ "mean_token_accuracy": 0.44971805810928345,
792
+ "num_tokens": 21880131.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 6.3291592128801435,
797
+ "grad_norm": 1.6503797769546509,
798
+ "learning_rate": 7.352517985611511e-06,
799
+ "loss": 2.951,
800
+ "mean_token_accuracy": 0.4522330954670906,
801
+ "num_tokens": 22131666.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 6.400715563506261,
806
+ "grad_norm": 1.1257638931274414,
807
+ "learning_rate": 7.20863309352518e-06,
808
+ "loss": 2.9598,
809
+ "mean_token_accuracy": 0.4505872465670109,
810
+ "num_tokens": 22385786.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 6.47227191413238,
815
+ "grad_norm": 0.5459461808204651,
816
+ "learning_rate": 7.064748201438849e-06,
817
+ "loss": 2.927,
818
+ "mean_token_accuracy": 0.45370263084769247,
819
+ "num_tokens": 22638350.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 6.543828264758497,
824
+ "grad_norm": 0.5412788987159729,
825
+ "learning_rate": 6.920863309352519e-06,
826
+ "loss": 2.9412,
827
+ "mean_token_accuracy": 0.4531597658991814,
828
+ "num_tokens": 22891362.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 6.615384615384615,
833
+ "grad_norm": 0.7567277550697327,
834
+ "learning_rate": 6.776978417266188e-06,
835
+ "loss": 2.9678,
836
+ "mean_token_accuracy": 0.449733317643404,
837
+ "num_tokens": 23140429.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 6.6869409660107335,
842
+ "grad_norm": 0.5896158814430237,
843
+ "learning_rate": 6.633093525179857e-06,
844
+ "loss": 2.9374,
845
+ "mean_token_accuracy": 0.45310442596673967,
846
+ "num_tokens": 23390979.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 6.758497316636851,
851
+ "grad_norm": 0.6928054690361023,
852
+ "learning_rate": 6.489208633093526e-06,
853
+ "loss": 2.9311,
854
+ "mean_token_accuracy": 0.4540563628077507,
855
+ "num_tokens": 23643194.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 6.83005366726297,
860
+ "grad_norm": 0.5281582474708557,
861
+ "learning_rate": 6.345323741007194e-06,
862
+ "loss": 2.931,
863
+ "mean_token_accuracy": 0.45377019196748736,
864
+ "num_tokens": 23896338.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 6.901610017889087,
869
+ "grad_norm": 0.6179826259613037,
870
+ "learning_rate": 6.2014388489208645e-06,
871
+ "loss": 2.9377,
872
+ "mean_token_accuracy": 0.45448167622089386,
873
+ "num_tokens": 24143728.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 6.973166368515206,
878
+ "grad_norm": 1.0639809370040894,
879
+ "learning_rate": 6.057553956834533e-06,
880
+ "loss": 2.9194,
881
+ "mean_token_accuracy": 0.45466252863407136,
882
+ "num_tokens": 24394294.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 7.0500894454382825,
887
+ "grad_norm": 1.7252315282821655,
888
+ "learning_rate": 5.913669064748202e-06,
889
+ "loss": 3.2321,
890
+ "mean_token_accuracy": 0.4523723830537098,
891
+ "num_tokens": 24654875.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 7.121645796064401,
896
+ "grad_norm": 2.1776368618011475,
897
+ "learning_rate": 5.769784172661871e-06,
898
+ "loss": 2.9234,
899
+ "mean_token_accuracy": 0.4544162482023239,
900
+ "num_tokens": 24908270.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 7.193202146690519,
905
+ "grad_norm": 0.46500054001808167,
906
+ "learning_rate": 5.62589928057554e-06,
907
+ "loss": 2.939,
908
+ "mean_token_accuracy": 0.4552768111228943,
909
+ "num_tokens": 25154564.0,
910
+ "step": 1000
911
+ },
912
+ {
913
+ "epoch": 7.264758497316637,
914
+ "grad_norm": 1.2805962562561035,
915
+ "learning_rate": 5.482014388489208e-06,
916
+ "loss": 2.9088,
917
+ "mean_token_accuracy": 0.4569986887276173,
918
+ "num_tokens": 25405656.0,
919
+ "step": 1010
920
+ },
921
+ {
922
+ "epoch": 7.336314847942755,
923
+ "grad_norm": 0.7547609806060791,
924
+ "learning_rate": 5.3381294964028786e-06,
925
+ "loss": 2.9138,
926
+ "mean_token_accuracy": 0.4570966839790344,
927
+ "num_tokens": 25650632.0,
928
+ "step": 1020
929
+ },
930
+ {
931
+ "epoch": 7.407871198568873,
932
+ "grad_norm": 0.5120713710784912,
933
+ "learning_rate": 5.194244604316547e-06,
934
+ "loss": 2.8993,
935
+ "mean_token_accuracy": 0.4579989641904831,
936
+ "num_tokens": 25903102.0,
937
+ "step": 1030
938
+ },
939
+ {
940
+ "epoch": 7.479427549194991,
941
+ "grad_norm": 0.5636111497879028,
942
+ "learning_rate": 5.050359712230216e-06,
943
+ "loss": 2.9162,
944
+ "mean_token_accuracy": 0.4563060127198696,
945
+ "num_tokens": 26155748.0,
946
+ "step": 1040
947
+ },
948
+ {
949
+ "epoch": 7.550983899821109,
950
+ "grad_norm": 0.7809133529663086,
951
+ "learning_rate": 4.9064748201438856e-06,
952
+ "loss": 2.914,
953
+ "mean_token_accuracy": 0.4561266407370567,
954
+ "num_tokens": 26405583.0,
955
+ "step": 1050
956
+ },
957
+ {
958
+ "epoch": 7.622540250447227,
959
+ "grad_norm": 0.4882202744483948,
960
+ "learning_rate": 4.762589928057554e-06,
961
+ "loss": 2.9361,
962
+ "mean_token_accuracy": 0.4527492173016071,
963
+ "num_tokens": 26655356.0,
964
+ "step": 1060
965
+ },
966
+ {
967
+ "epoch": 7.694096601073345,
968
+ "grad_norm": 0.46441176533699036,
969
+ "learning_rate": 4.618705035971223e-06,
970
+ "loss": 2.9233,
971
+ "mean_token_accuracy": 0.45422032177448274,
972
+ "num_tokens": 26911716.0,
973
+ "step": 1070
974
+ },
975
+ {
976
+ "epoch": 7.7656529516994635,
977
+ "grad_norm": 0.6556686162948608,
978
+ "learning_rate": 4.474820143884893e-06,
979
+ "loss": 2.9213,
980
+ "mean_token_accuracy": 0.45541969388723375,
981
+ "num_tokens": 27160591.0,
982
+ "step": 1080
983
+ },
984
+ {
985
+ "epoch": 7.837209302325581,
986
+ "grad_norm": 1.0379059314727783,
987
+ "learning_rate": 4.330935251798561e-06,
988
+ "loss": 2.8881,
989
+ "mean_token_accuracy": 0.4577357158064842,
990
+ "num_tokens": 27414873.0,
991
+ "step": 1090
992
+ },
993
+ {
994
+ "epoch": 7.9087656529517,
995
+ "grad_norm": 0.8527312874794006,
996
+ "learning_rate": 4.18705035971223e-06,
997
+ "loss": 2.8882,
998
+ "mean_token_accuracy": 0.46129798144102097,
999
+ "num_tokens": 27663071.0,
1000
+ "step": 1100
1001
+ },
1002
+ {
1003
+ "epoch": 7.980322003577817,
1004
+ "grad_norm": 0.4619864225387573,
1005
+ "learning_rate": 4.0431654676259e-06,
1006
+ "loss": 2.932,
1007
+ "mean_token_accuracy": 0.4540557064116001,
1008
+ "num_tokens": 27916696.0,
1009
+ "step": 1110
1010
+ },
1011
+ {
1012
+ "epoch": 8.057245080500895,
1013
+ "grad_norm": 0.7810168266296387,
1014
+ "learning_rate": 3.899280575539569e-06,
1015
+ "loss": 3.1952,
1016
+ "mean_token_accuracy": 0.4580523044597812,
1017
+ "num_tokens": 28170819.0,
1018
+ "step": 1120
1019
+ },
1020
+ {
1021
+ "epoch": 8.128801431127012,
1022
+ "grad_norm": 0.5160698890686035,
1023
+ "learning_rate": 3.7553956834532378e-06,
1024
+ "loss": 2.8928,
1025
+ "mean_token_accuracy": 0.45813094154000283,
1026
+ "num_tokens": 28424784.0,
1027
+ "step": 1130
1028
+ },
1029
+ {
1030
+ "epoch": 8.20035778175313,
1031
+ "grad_norm": 0.44342929124832153,
1032
+ "learning_rate": 3.6115107913669066e-06,
1033
+ "loss": 2.9013,
1034
+ "mean_token_accuracy": 0.4571513183414936,
1035
+ "num_tokens": 28675178.0,
1036
+ "step": 1140
1037
+ },
1038
+ {
1039
+ "epoch": 8.271914132379248,
1040
+ "grad_norm": 0.6888077855110168,
1041
+ "learning_rate": 3.4676258992805755e-06,
1042
+ "loss": 2.9162,
1043
+ "mean_token_accuracy": 0.45669827088713644,
1044
+ "num_tokens": 28924005.0,
1045
+ "step": 1150
1046
+ },
1047
+ {
1048
+ "epoch": 8.343470483005367,
1049
+ "grad_norm": 0.5610206127166748,
1050
+ "learning_rate": 3.323741007194245e-06,
1051
+ "loss": 2.9215,
1052
+ "mean_token_accuracy": 0.45746611058712006,
1053
+ "num_tokens": 29175601.0,
1054
+ "step": 1160
1055
+ },
1056
+ {
1057
+ "epoch": 8.415026833631485,
1058
+ "grad_norm": 0.4491877555847168,
1059
+ "learning_rate": 3.1798561151079137e-06,
1060
+ "loss": 2.8979,
1061
+ "mean_token_accuracy": 0.4579362317919731,
1062
+ "num_tokens": 29424141.0,
1063
+ "step": 1170
1064
+ },
1065
+ {
1066
+ "epoch": 8.486583184257602,
1067
+ "grad_norm": 0.7424222230911255,
1068
+ "learning_rate": 3.0359712230215834e-06,
1069
+ "loss": 2.8942,
1070
+ "mean_token_accuracy": 0.4580856651067734,
1071
+ "num_tokens": 29681514.0,
1072
+ "step": 1180
1073
+ },
1074
+ {
1075
+ "epoch": 8.55813953488372,
1076
+ "grad_norm": 1.396942138671875,
1077
+ "learning_rate": 2.892086330935252e-06,
1078
+ "loss": 2.9007,
1079
+ "mean_token_accuracy": 0.4569838471710682,
1080
+ "num_tokens": 29932547.0,
1081
+ "step": 1190
1082
+ },
1083
+ {
1084
+ "epoch": 8.62969588550984,
1085
+ "grad_norm": 0.43744152784347534,
1086
+ "learning_rate": 2.7482014388489207e-06,
1087
+ "loss": 2.8771,
1088
+ "mean_token_accuracy": 0.46146027445793153,
1089
+ "num_tokens": 30185906.0,
1090
+ "step": 1200
1091
+ },
1092
+ {
1093
+ "epoch": 8.701252236135957,
1094
+ "grad_norm": 2.9691150188446045,
1095
+ "learning_rate": 2.6043165467625904e-06,
1096
+ "loss": 2.8739,
1097
+ "mean_token_accuracy": 0.4621762208640575,
1098
+ "num_tokens": 30431346.0,
1099
+ "step": 1210
1100
+ },
1101
+ {
1102
+ "epoch": 8.772808586762075,
1103
+ "grad_norm": 0.5249901413917542,
1104
+ "learning_rate": 2.4604316546762593e-06,
1105
+ "loss": 2.9069,
1106
+ "mean_token_accuracy": 0.4558020316064358,
1107
+ "num_tokens": 30680282.0,
1108
+ "step": 1220
1109
+ },
1110
+ {
1111
+ "epoch": 8.844364937388193,
1112
+ "grad_norm": 0.8537925481796265,
1113
+ "learning_rate": 2.316546762589928e-06,
1114
+ "loss": 2.897,
1115
+ "mean_token_accuracy": 0.4572525009512901,
1116
+ "num_tokens": 30935075.0,
1117
+ "step": 1230
1118
+ },
1119
+ {
1120
+ "epoch": 8.915921288014312,
1121
+ "grad_norm": 0.5540444850921631,
1122
+ "learning_rate": 2.1726618705035974e-06,
1123
+ "loss": 2.8822,
1124
+ "mean_token_accuracy": 0.46007697507739065,
1125
+ "num_tokens": 31183768.0,
1126
+ "step": 1240
1127
+ },
1128
+ {
1129
+ "epoch": 8.98747763864043,
1130
+ "grad_norm": 0.6158258318901062,
1131
+ "learning_rate": 2.0287769784172663e-06,
1132
+ "loss": 2.8853,
1133
+ "mean_token_accuracy": 0.45986568853259085,
1134
+ "num_tokens": 31436091.0,
1135
+ "step": 1250
1136
+ },
1137
+ {
1138
+ "epoch": 9.064400715563506,
1139
+ "grad_norm": 4.425424098968506,
1140
+ "learning_rate": 1.8848920863309354e-06,
1141
+ "loss": 3.1999,
1142
+ "mean_token_accuracy": 0.45752188490658274,
1143
+ "num_tokens": 31693958.0,
1144
+ "step": 1260
1145
+ },
1146
+ {
1147
+ "epoch": 9.135957066189624,
1148
+ "grad_norm": 0.6236605048179626,
1149
+ "learning_rate": 1.7410071942446045e-06,
1150
+ "loss": 2.8925,
1151
+ "mean_token_accuracy": 0.4580650545656681,
1152
+ "num_tokens": 31943050.0,
1153
+ "step": 1270
1154
+ },
1155
+ {
1156
+ "epoch": 9.207513416815742,
1157
+ "grad_norm": 0.5942525863647461,
1158
+ "learning_rate": 1.5971223021582735e-06,
1159
+ "loss": 2.8896,
1160
+ "mean_token_accuracy": 0.45962206125259397,
1161
+ "num_tokens": 32193786.0,
1162
+ "step": 1280
1163
+ },
1164
+ {
1165
+ "epoch": 9.279069767441861,
1166
+ "grad_norm": 0.9935563802719116,
1167
+ "learning_rate": 1.4532374100719426e-06,
1168
+ "loss": 2.8982,
1169
+ "mean_token_accuracy": 0.457182715088129,
1170
+ "num_tokens": 32449326.0,
1171
+ "step": 1290
1172
+ },
1173
+ {
1174
+ "epoch": 9.350626118067979,
1175
+ "grad_norm": 0.8188269734382629,
1176
+ "learning_rate": 1.3093525179856117e-06,
1177
+ "loss": 2.855,
1178
+ "mean_token_accuracy": 0.46316581666469575,
1179
+ "num_tokens": 32698038.0,
1180
+ "step": 1300
1181
+ },
1182
+ {
1183
+ "epoch": 9.422182468694096,
1184
+ "grad_norm": 0.6339349150657654,
1185
+ "learning_rate": 1.1654676258992808e-06,
1186
+ "loss": 2.873,
1187
+ "mean_token_accuracy": 0.4594662845134735,
1188
+ "num_tokens": 32949656.0,
1189
+ "step": 1310
1190
+ },
1191
+ {
1192
+ "epoch": 9.493738819320214,
1193
+ "grad_norm": 0.6287016868591309,
1194
+ "learning_rate": 1.0215827338129496e-06,
1195
+ "loss": 2.8851,
1196
+ "mean_token_accuracy": 0.45849224627017976,
1197
+ "num_tokens": 33203989.0,
1198
+ "step": 1320
1199
+ },
1200
+ {
1201
+ "epoch": 9.565295169946333,
1202
+ "grad_norm": 0.6148714423179626,
1203
+ "learning_rate": 8.776978417266188e-07,
1204
+ "loss": 2.9034,
1205
+ "mean_token_accuracy": 0.4570117011666298,
1206
+ "num_tokens": 33456115.0,
1207
+ "step": 1330
1208
+ },
1209
+ {
1210
+ "epoch": 9.636851520572451,
1211
+ "grad_norm": 0.5148676633834839,
1212
+ "learning_rate": 7.338129496402878e-07,
1213
+ "loss": 2.892,
1214
+ "mean_token_accuracy": 0.459391076117754,
1215
+ "num_tokens": 33711492.0,
1216
+ "step": 1340
1217
+ },
1218
+ {
1219
+ "epoch": 9.708407871198569,
1220
+ "grad_norm": 0.9138558506965637,
1221
+ "learning_rate": 5.899280575539569e-07,
1222
+ "loss": 2.8813,
1223
+ "mean_token_accuracy": 0.46166028156876565,
1224
+ "num_tokens": 33958698.0,
1225
+ "step": 1350
1226
+ },
1227
+ {
1228
+ "epoch": 9.779964221824686,
1229
+ "grad_norm": 0.7092456817626953,
1230
+ "learning_rate": 4.4604316546762596e-07,
1231
+ "loss": 2.8844,
1232
+ "mean_token_accuracy": 0.4608030468225479,
1233
+ "num_tokens": 34205747.0,
1234
+ "step": 1360
1235
+ },
1236
+ {
1237
+ "epoch": 9.851520572450806,
1238
+ "grad_norm": 0.8316847681999207,
1239
+ "learning_rate": 3.02158273381295e-07,
1240
+ "loss": 2.9008,
1241
+ "mean_token_accuracy": 0.4585262954235077,
1242
+ "num_tokens": 34458362.0,
1243
+ "step": 1370
1244
+ },
1245
+ {
1246
+ "epoch": 9.923076923076923,
1247
+ "grad_norm": 0.6987820267677307,
1248
+ "learning_rate": 1.5827338129496404e-07,
1249
+ "loss": 2.8819,
1250
+ "mean_token_accuracy": 0.4603422984480858,
1251
+ "num_tokens": 34708914.0,
1252
+ "step": 1380
1253
+ },
1254
+ {
1255
+ "epoch": 9.994633273703041,
1256
+ "grad_norm": 0.7729736566543579,
1257
+ "learning_rate": 1.4388489208633095e-08,
1258
+ "loss": 2.9041,
1259
+ "mean_token_accuracy": 0.45747644156217576,
1260
+ "num_tokens": 34959957.0,
1261
+ "step": 1390
1262
+ }
1263
+ ],
1264
+ "logging_steps": 10,
1265
+ "max_steps": 1390,
1266
+ "num_input_tokens_seen": 0,
1267
+ "num_train_epochs": 10,
1268
+ "save_steps": 500,
1269
+ "stateful_callbacks": {
1270
+ "TrainerControl": {
1271
+ "args": {
1272
+ "should_epoch_stop": false,
1273
+ "should_evaluate": false,
1274
+ "should_log": false,
1275
+ "should_save": true,
1276
+ "should_training_stop": true
1277
+ },
1278
+ "attributes": {}
1279
+ }
1280
+ },
1281
+ "total_flos": 3.858756252903014e+16,
1282
+ "train_batch_size": 64,
1283
+ "trial_name": null,
1284
+ "trial_params": null
1285
+ }
checkpoint-1390/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d25b53cc47734bcef31e00edf7734c0385193e3d581c4ba0c2cd909723a664bb
3
+ size 5688
checkpoint-500/config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 1024,
14
+ "max_position_embeddings": 512,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 6,
18
+ "num_hidden_layers": 6,
19
+ "num_key_value_heads": 6,
20
+ "pad_token_id": 0,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-05,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.51.3",
28
+ "use_cache": true,
29
+ "vocab_size": 32000
30
+ }
checkpoint-500/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.51.3"
7
+ }
checkpoint-500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:366dd207462f94c865b13bbffc940a448332d52ea7411652cd817b9f4396462d
3
+ size 309900448
checkpoint-500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f71b4798f043306ba9a764799fc6ae83a2ba22324528263502da41564940b60
3
+ size 619836730
checkpoint-500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f37c40ce327861a7ca13b719d3aa37510a143368b6e74358bdb14becb3899e1e
3
+ size 14244
checkpoint-500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ab2c291cab7c28fd90ff222c6ef90ee4060b3d3475da1afe0a061fd24fb16f5
3
+ size 1064
checkpoint-500/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
checkpoint-500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-500/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-500/tokenizer_config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": true,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": true,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "extra_special_tokens": {},
35
+ "legacy": false,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "sp_model_kwargs": {},
39
+ "tokenizer_class": "LlamaTokenizer",
40
+ "unk_token": "<unk>",
41
+ "use_default_system_prompt": false
42
+ }
checkpoint-500/trainer_state.json ADDED
@@ -0,0 +1,484 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.59391771019678,
6
+ "eval_steps": 500,
7
+ "global_step": 500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.07155635062611806,
14
+ "grad_norm": 3.981292247772217,
15
+ "learning_rate": 1.9870503597122304e-05,
16
+ "loss": 6.5969,
17
+ "mean_token_accuracy": 0.15720859691500663,
18
+ "num_tokens": 249014.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.14311270125223613,
23
+ "grad_norm": 6.078427314758301,
24
+ "learning_rate": 1.9726618705035972e-05,
25
+ "loss": 6.3047,
26
+ "mean_token_accuracy": 0.17161023132503034,
27
+ "num_tokens": 499527.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.2146690518783542,
32
+ "grad_norm": 3.555305242538452,
33
+ "learning_rate": 1.958273381294964e-05,
34
+ "loss": 6.0472,
35
+ "mean_token_accuracy": 0.19869447350502015,
36
+ "num_tokens": 755405.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.28622540250447226,
41
+ "grad_norm": 3.784306764602661,
42
+ "learning_rate": 1.9438848920863313e-05,
43
+ "loss": 5.8397,
44
+ "mean_token_accuracy": 0.22695979550480844,
45
+ "num_tokens": 1011193.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.35778175313059035,
50
+ "grad_norm": 3.2327146530151367,
51
+ "learning_rate": 1.9294964028776978e-05,
52
+ "loss": 5.6695,
53
+ "mean_token_accuracy": 0.2401440929621458,
54
+ "num_tokens": 1258965.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.4293381037567084,
59
+ "grad_norm": 3.420456647872925,
60
+ "learning_rate": 1.915107913669065e-05,
61
+ "loss": 5.4321,
62
+ "mean_token_accuracy": 0.2521901123225689,
63
+ "num_tokens": 1512466.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.5008944543828264,
68
+ "grad_norm": 4.493953227996826,
69
+ "learning_rate": 1.9007194244604318e-05,
70
+ "loss": 5.2609,
71
+ "mean_token_accuracy": 0.26822467111051085,
72
+ "num_tokens": 1760599.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.5724508050089445,
77
+ "grad_norm": 2.4932243824005127,
78
+ "learning_rate": 1.8863309352517986e-05,
79
+ "loss": 5.1328,
80
+ "mean_token_accuracy": 0.275511983782053,
81
+ "num_tokens": 2014670.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.6440071556350626,
86
+ "grad_norm": 2.402729034423828,
87
+ "learning_rate": 1.8719424460431658e-05,
88
+ "loss": 5.0135,
89
+ "mean_token_accuracy": 0.2811047907918692,
90
+ "num_tokens": 2264498.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.7155635062611807,
95
+ "grad_norm": 2.8561453819274902,
96
+ "learning_rate": 1.8575539568345323e-05,
97
+ "loss": 4.8244,
98
+ "mean_token_accuracy": 0.2967699430882931,
99
+ "num_tokens": 2508975.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.7871198568872988,
104
+ "grad_norm": 2.222463846206665,
105
+ "learning_rate": 1.8431654676258995e-05,
106
+ "loss": 4.7404,
107
+ "mean_token_accuracy": 0.30234326496720315,
108
+ "num_tokens": 2760092.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.8586762075134168,
113
+ "grad_norm": 2.0150718688964844,
114
+ "learning_rate": 1.8287769784172664e-05,
115
+ "loss": 4.5925,
116
+ "mean_token_accuracy": 0.31090952083468437,
117
+ "num_tokens": 3011643.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.9302325581395349,
122
+ "grad_norm": 2.499884605407715,
123
+ "learning_rate": 1.8143884892086332e-05,
124
+ "loss": 4.484,
125
+ "mean_token_accuracy": 0.3174323089420795,
126
+ "num_tokens": 3260597.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 1.0071556350626119,
131
+ "grad_norm": 3.81246280670166,
132
+ "learning_rate": 1.8e-05,
133
+ "loss": 4.8272,
134
+ "mean_token_accuracy": 0.32652662367355534,
135
+ "num_tokens": 3521468.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 1.07871198568873,
140
+ "grad_norm": 1.3433386087417603,
141
+ "learning_rate": 1.785611510791367e-05,
142
+ "loss": 4.2527,
143
+ "mean_token_accuracy": 0.3359356351196766,
144
+ "num_tokens": 3764868.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 1.150268336314848,
149
+ "grad_norm": 2.1221790313720703,
150
+ "learning_rate": 1.771223021582734e-05,
151
+ "loss": 4.1608,
152
+ "mean_token_accuracy": 0.33773822858929636,
153
+ "num_tokens": 4013198.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 1.2218246869409661,
158
+ "grad_norm": 1.3530231714248657,
159
+ "learning_rate": 1.756834532374101e-05,
160
+ "loss": 4.0771,
161
+ "mean_token_accuracy": 0.3416416011750698,
162
+ "num_tokens": 4266761.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 1.293381037567084,
167
+ "grad_norm": 8.359506607055664,
168
+ "learning_rate": 1.7424460431654678e-05,
169
+ "loss": 3.9872,
170
+ "mean_token_accuracy": 0.3446764849126339,
171
+ "num_tokens": 4524185.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 1.364937388193202,
176
+ "grad_norm": 0.9995141625404358,
177
+ "learning_rate": 1.7280575539568346e-05,
178
+ "loss": 3.8942,
179
+ "mean_token_accuracy": 0.3535850510001183,
180
+ "num_tokens": 4775262.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 1.4364937388193202,
185
+ "grad_norm": 0.9777733087539673,
186
+ "learning_rate": 1.7136690647482015e-05,
187
+ "loss": 3.8149,
188
+ "mean_token_accuracy": 0.36338763535022733,
189
+ "num_tokens": 5022677.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 1.5080500894454383,
194
+ "grad_norm": 0.9398026466369629,
195
+ "learning_rate": 1.6992805755395686e-05,
196
+ "loss": 3.7274,
197
+ "mean_token_accuracy": 0.37123609259724616,
198
+ "num_tokens": 5271810.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 1.5796064400715564,
203
+ "grad_norm": 3.464630603790283,
204
+ "learning_rate": 1.6848920863309355e-05,
205
+ "loss": 3.7154,
206
+ "mean_token_accuracy": 0.3758062146604061,
207
+ "num_tokens": 5522408.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 1.6511627906976745,
212
+ "grad_norm": 2.104501485824585,
213
+ "learning_rate": 1.6705035971223023e-05,
214
+ "loss": 3.6469,
215
+ "mean_token_accuracy": 0.3832809627056122,
216
+ "num_tokens": 5769231.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 1.7227191413237923,
221
+ "grad_norm": 2.5066070556640625,
222
+ "learning_rate": 1.656115107913669e-05,
223
+ "loss": 3.6434,
224
+ "mean_token_accuracy": 0.38175718635320666,
225
+ "num_tokens": 6020808.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 1.7942754919499104,
230
+ "grad_norm": 0.9507914781570435,
231
+ "learning_rate": 1.641726618705036e-05,
232
+ "loss": 3.5866,
233
+ "mean_token_accuracy": 0.3858716659247875,
234
+ "num_tokens": 6275383.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 1.8658318425760285,
239
+ "grad_norm": 3.4407248497009277,
240
+ "learning_rate": 1.627338129496403e-05,
241
+ "loss": 3.5663,
242
+ "mean_token_accuracy": 0.38949630260467527,
243
+ "num_tokens": 6525182.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 1.9373881932021466,
248
+ "grad_norm": 0.7253276109695435,
249
+ "learning_rate": 1.61294964028777e-05,
250
+ "loss": 3.5233,
251
+ "mean_token_accuracy": 0.391743092238903,
252
+ "num_tokens": 6779756.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 2.0143112701252237,
257
+ "grad_norm": 0.6907745599746704,
258
+ "learning_rate": 1.598561151079137e-05,
259
+ "loss": 3.8364,
260
+ "mean_token_accuracy": 0.395021445140606,
261
+ "num_tokens": 7043995.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 2.085867620751342,
266
+ "grad_norm": 0.9332271218299866,
267
+ "learning_rate": 1.5841726618705037e-05,
268
+ "loss": 3.4772,
269
+ "mean_token_accuracy": 0.3965952657163143,
270
+ "num_tokens": 7297135.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 2.15742397137746,
275
+ "grad_norm": 1.235854983329773,
276
+ "learning_rate": 1.5697841726618706e-05,
277
+ "loss": 3.4414,
278
+ "mean_token_accuracy": 0.3989549733698368,
279
+ "num_tokens": 7545580.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 2.228980322003578,
284
+ "grad_norm": 0.6656942963600159,
285
+ "learning_rate": 1.5553956834532374e-05,
286
+ "loss": 3.4114,
287
+ "mean_token_accuracy": 0.40089740604162216,
288
+ "num_tokens": 7799266.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 2.300536672629696,
293
+ "grad_norm": 0.6185078024864197,
294
+ "learning_rate": 1.5410071942446046e-05,
295
+ "loss": 3.395,
296
+ "mean_token_accuracy": 0.4044420339167118,
297
+ "num_tokens": 8052210.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 2.3720930232558137,
302
+ "grad_norm": 1.1096031665802002,
303
+ "learning_rate": 1.5266187050359714e-05,
304
+ "loss": 3.3804,
305
+ "mean_token_accuracy": 0.40613374635577204,
306
+ "num_tokens": 8307281.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 2.4436493738819323,
311
+ "grad_norm": 0.5653504729270935,
312
+ "learning_rate": 1.5122302158273383e-05,
313
+ "loss": 3.3501,
314
+ "mean_token_accuracy": 0.40836238488554955,
315
+ "num_tokens": 8557428.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 2.51520572450805,
320
+ "grad_norm": 0.5432367920875549,
321
+ "learning_rate": 1.4978417266187051e-05,
322
+ "loss": 3.3467,
323
+ "mean_token_accuracy": 0.40912329107522966,
324
+ "num_tokens": 8808315.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 2.586762075134168,
329
+ "grad_norm": 1.6327078342437744,
330
+ "learning_rate": 1.4834532374100721e-05,
331
+ "loss": 3.3258,
332
+ "mean_token_accuracy": 0.4106706023216248,
333
+ "num_tokens": 9057886.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 2.658318425760286,
338
+ "grad_norm": 0.6323248147964478,
339
+ "learning_rate": 1.469064748201439e-05,
340
+ "loss": 3.298,
341
+ "mean_token_accuracy": 0.41263881102204325,
342
+ "num_tokens": 9309357.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 2.729874776386404,
347
+ "grad_norm": 2.182023525238037,
348
+ "learning_rate": 1.4546762589928058e-05,
349
+ "loss": 3.2819,
350
+ "mean_token_accuracy": 0.4163250602781773,
351
+ "num_tokens": 9558445.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 2.8014311270125223,
356
+ "grad_norm": 0.5517727136611938,
357
+ "learning_rate": 1.4402877697841728e-05,
358
+ "loss": 3.2486,
359
+ "mean_token_accuracy": 0.4189229063689709,
360
+ "num_tokens": 9813513.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 2.8729874776386404,
365
+ "grad_norm": 0.877505362033844,
366
+ "learning_rate": 1.4258992805755397e-05,
367
+ "loss": 3.2428,
368
+ "mean_token_accuracy": 0.4193955764174461,
369
+ "num_tokens": 10059533.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 2.9445438282647585,
374
+ "grad_norm": 0.7058685421943665,
375
+ "learning_rate": 1.4115107913669065e-05,
376
+ "loss": 3.2369,
377
+ "mean_token_accuracy": 0.4203347131609917,
378
+ "num_tokens": 10309900.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 3.0214669051878356,
383
+ "grad_norm": 1.2357401847839355,
384
+ "learning_rate": 1.3971223021582736e-05,
385
+ "loss": 3.563,
386
+ "mean_token_accuracy": 0.4190788792400825,
387
+ "num_tokens": 10568955.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 3.0930232558139537,
392
+ "grad_norm": 0.7918577790260315,
393
+ "learning_rate": 1.3827338129496404e-05,
394
+ "loss": 3.2081,
395
+ "mean_token_accuracy": 0.4243326403200626,
396
+ "num_tokens": 10818378.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 3.1645796064400717,
401
+ "grad_norm": 0.5052043795585632,
402
+ "learning_rate": 1.3683453237410074e-05,
403
+ "loss": 3.2022,
404
+ "mean_token_accuracy": 0.4242256350815296,
405
+ "num_tokens": 11070954.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 3.23613595706619,
410
+ "grad_norm": 0.5182781219482422,
411
+ "learning_rate": 1.353956834532374e-05,
412
+ "loss": 3.199,
413
+ "mean_token_accuracy": 0.4230089582502842,
414
+ "num_tokens": 11323582.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 3.3076923076923075,
419
+ "grad_norm": 1.1913461685180664,
420
+ "learning_rate": 1.3395683453237411e-05,
421
+ "loss": 3.1853,
422
+ "mean_token_accuracy": 0.42622370198369025,
423
+ "num_tokens": 11574310.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 3.3792486583184256,
428
+ "grad_norm": 3.5984549522399902,
429
+ "learning_rate": 1.3251798561151081e-05,
430
+ "loss": 3.1743,
431
+ "mean_token_accuracy": 0.42716887295246125,
432
+ "num_tokens": 11828106.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 3.4508050089445437,
437
+ "grad_norm": 0.8405038118362427,
438
+ "learning_rate": 1.310791366906475e-05,
439
+ "loss": 3.1591,
440
+ "mean_token_accuracy": 0.4280025988817215,
441
+ "num_tokens": 12078899.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 3.5223613595706618,
446
+ "grad_norm": 0.5538778901100159,
447
+ "learning_rate": 1.296402877697842e-05,
448
+ "loss": 3.1318,
449
+ "mean_token_accuracy": 0.43028812259435656,
450
+ "num_tokens": 12331071.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 3.59391771019678,
455
+ "grad_norm": 2.8024959564208984,
456
+ "learning_rate": 1.2820143884892086e-05,
457
+ "loss": 3.1394,
458
+ "mean_token_accuracy": 0.4312668785452843,
459
+ "num_tokens": 12581208.0,
460
+ "step": 500
461
+ }
462
+ ],
463
+ "logging_steps": 10,
464
+ "max_steps": 1390,
465
+ "num_input_tokens_seen": 0,
466
+ "num_train_epochs": 10,
467
+ "save_steps": 500,
468
+ "stateful_callbacks": {
469
+ "TrainerControl": {
470
+ "args": {
471
+ "should_epoch_stop": false,
472
+ "should_evaluate": false,
473
+ "should_log": false,
474
+ "should_save": true,
475
+ "should_training_stop": false
476
+ },
477
+ "attributes": {}
478
+ }
479
+ },
480
+ "total_flos": 1.3969337037225984e+16,
481
+ "train_batch_size": 64,
482
+ "trial_name": null,
483
+ "trial_params": null
484
+ }
checkpoint-500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d25b53cc47734bcef31e00edf7734c0385193e3d581c4ba0c2cd909723a664bb
3
+ size 5688
config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 1024,
14
+ "max_position_embeddings": 512,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 6,
18
+ "num_hidden_layers": 6,
19
+ "num_key_value_heads": 6,
20
+ "pad_token_id": 0,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-05,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.51.3",
28
+ "use_cache": true,
29
+ "vocab_size": 32000
30
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.51.3"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc1ecc51f6ae0b74b7232e22a94a71d7ffb72ce57a924be13dc83c8af25c6af6
3
+ size 309900448
runs/Apr25_07-50-36_9c18e885a673/events.out.tfevents.1745567454.9c18e885a673.5280.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9496644bfd07f191de4fe1f4e87267f7f920865d707bead83b391dd15036a1c8
3
+ size 51641
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": true,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": true,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "extra_special_tokens": {},
35
+ "legacy": false,
36
+ "model_max_length": 1000000000000000019884624838656,
37
+ "pad_token": "</s>",
38
+ "sp_model_kwargs": {},
39
+ "tokenizer_class": "LlamaTokenizer",
40
+ "unk_token": "<unk>",
41
+ "use_default_system_prompt": false
42
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d25b53cc47734bcef31e00edf7734c0385193e3d581c4ba0c2cd909723a664bb
3
+ size 5688