rootxhacker commited on
Commit
d7bfc08
·
verified ·
1 Parent(s): 909b509

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. 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-1500/config.json +30 -0
  14. checkpoint-1500/generation_config.json +7 -0
  15. checkpoint-1500/model.safetensors +3 -0
  16. checkpoint-1500/optimizer.pt +3 -0
  17. checkpoint-1500/rng_state.pth +3 -0
  18. checkpoint-1500/scheduler.pt +3 -0
  19. checkpoint-1500/special_tokens_map.json +24 -0
  20. checkpoint-1500/tokenizer.json +0 -0
  21. checkpoint-1500/tokenizer.model +3 -0
  22. checkpoint-1500/tokenizer_config.json +42 -0
  23. checkpoint-1500/trainer_state.json +1384 -0
  24. checkpoint-1500/training_args.bin +3 -0
  25. checkpoint-2000/config.json +30 -0
  26. checkpoint-2000/generation_config.json +7 -0
  27. checkpoint-2000/model.safetensors +3 -0
  28. checkpoint-2000/optimizer.pt +3 -0
  29. checkpoint-2000/rng_state.pth +3 -0
  30. checkpoint-2000/scheduler.pt +3 -0
  31. checkpoint-2000/special_tokens_map.json +24 -0
  32. checkpoint-2000/tokenizer.json +0 -0
  33. checkpoint-2000/tokenizer.model +3 -0
  34. checkpoint-2000/tokenizer_config.json +42 -0
  35. checkpoint-2000/trainer_state.json +1834 -0
  36. checkpoint-2000/training_args.bin +3 -0
  37. checkpoint-2500/config.json +30 -0
  38. checkpoint-2500/generation_config.json +7 -0
  39. checkpoint-2500/model.safetensors +3 -0
  40. checkpoint-2500/optimizer.pt +3 -0
  41. checkpoint-2500/rng_state.pth +3 -0
  42. checkpoint-2500/scheduler.pt +3 -0
  43. checkpoint-2500/special_tokens_map.json +24 -0
  44. checkpoint-2500/tokenizer.json +0 -0
  45. checkpoint-2500/tokenizer.model +3 -0
  46. checkpoint-2500/tokenizer_config.json +42 -0
  47. checkpoint-2500/trainer_state.json +2284 -0
  48. checkpoint-2500/training_args.bin +3 -0
  49. checkpoint-3000/config.json +30 -0
  50. checkpoint-3000/generation_config.json +7 -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": 256,
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 sha256:2b8e0592fed9f74babab260dfdd114ff456ba2379eb40d7c966dac20982c1772
3
+ size 309900448
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fb39655d8d5722a46fa244bf3b95648deac8b240add6ad2ad8428a7a0818dd4
3
+ size 619836730
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f07d6cf5a7ff917f31bde9a47ebed3e27fd53ff4b7099f3be0417151329a1f8
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": 0.3091190108191654,
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": 0.0030911901081916537,
14
+ "grad_norm": 3.058769941329956,
15
+ "learning_rate": 1.9972179289026277e-05,
16
+ "loss": 5.0209,
17
+ "mean_token_accuracy": 0.21833103336393833,
18
+ "num_tokens": 28941.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.0061823802163833074,
23
+ "grad_norm": 2.485980749130249,
24
+ "learning_rate": 1.994126738794436e-05,
25
+ "loss": 4.7631,
26
+ "mean_token_accuracy": 0.23807235918939113,
27
+ "num_tokens": 60990.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.00927357032457496,
32
+ "grad_norm": 5.288544178009033,
33
+ "learning_rate": 1.9910355486862444e-05,
34
+ "loss": 4.6799,
35
+ "mean_token_accuracy": 0.24858475103974342,
36
+ "num_tokens": 90602.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.012364760432766615,
41
+ "grad_norm": 3.226719379425049,
42
+ "learning_rate": 1.9879443585780528e-05,
43
+ "loss": 4.5883,
44
+ "mean_token_accuracy": 0.2555678006261587,
45
+ "num_tokens": 121132.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.015455950540958269,
50
+ "grad_norm": 2.027597188949585,
51
+ "learning_rate": 1.984853168469861e-05,
52
+ "loss": 4.5125,
53
+ "mean_token_accuracy": 0.26637452803552153,
54
+ "num_tokens": 154244.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.01854714064914992,
59
+ "grad_norm": 1.6715962886810303,
60
+ "learning_rate": 1.9817619783616695e-05,
61
+ "loss": 4.4224,
62
+ "mean_token_accuracy": 0.27198897041380404,
63
+ "num_tokens": 185618.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.021638330757341576,
68
+ "grad_norm": 3.0472540855407715,
69
+ "learning_rate": 1.9786707882534775e-05,
70
+ "loss": 4.3464,
71
+ "mean_token_accuracy": 0.2801041007041931,
72
+ "num_tokens": 220138.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.02472952086553323,
77
+ "grad_norm": 2.74045729637146,
78
+ "learning_rate": 1.9755795981452862e-05,
79
+ "loss": 4.4569,
80
+ "mean_token_accuracy": 0.2733523309230804,
81
+ "num_tokens": 252919.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.027820710973724884,
86
+ "grad_norm": 1.9274535179138184,
87
+ "learning_rate": 1.9724884080370946e-05,
88
+ "loss": 4.2569,
89
+ "mean_token_accuracy": 0.2858310595154762,
90
+ "num_tokens": 284739.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.030911901081916538,
95
+ "grad_norm": 1.906964898109436,
96
+ "learning_rate": 1.9693972179289026e-05,
97
+ "loss": 4.2144,
98
+ "mean_token_accuracy": 0.2953581381589174,
99
+ "num_tokens": 315186.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.03400309119010819,
104
+ "grad_norm": 2.042823314666748,
105
+ "learning_rate": 1.966306027820711e-05,
106
+ "loss": 4.2595,
107
+ "mean_token_accuracy": 0.287172843888402,
108
+ "num_tokens": 347507.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.03709428129829984,
113
+ "grad_norm": 1.9889057874679565,
114
+ "learning_rate": 1.9632148377125197e-05,
115
+ "loss": 4.1643,
116
+ "mean_token_accuracy": 0.30306958928704264,
117
+ "num_tokens": 380331.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.0401854714064915,
122
+ "grad_norm": 1.7952933311462402,
123
+ "learning_rate": 1.9601236476043277e-05,
124
+ "loss": 4.1334,
125
+ "mean_token_accuracy": 0.3080880597233772,
126
+ "num_tokens": 410947.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.04327666151468315,
131
+ "grad_norm": 3.8508658409118652,
132
+ "learning_rate": 1.957032457496136e-05,
133
+ "loss": 4.0262,
134
+ "mean_token_accuracy": 0.3118089348077774,
135
+ "num_tokens": 444130.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.04636785162287481,
140
+ "grad_norm": 31.280546188354492,
141
+ "learning_rate": 1.9539412673879444e-05,
142
+ "loss": 3.9889,
143
+ "mean_token_accuracy": 0.32304659858345985,
144
+ "num_tokens": 474570.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.04945904173106646,
149
+ "grad_norm": 2.1569416522979736,
150
+ "learning_rate": 1.950850077279753e-05,
151
+ "loss": 4.1053,
152
+ "mean_token_accuracy": 0.31258094161748884,
153
+ "num_tokens": 505047.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.05255023183925812,
158
+ "grad_norm": 1.808377981185913,
159
+ "learning_rate": 1.947758887171561e-05,
160
+ "loss": 3.9711,
161
+ "mean_token_accuracy": 0.33312112018465995,
162
+ "num_tokens": 533187.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.05564142194744977,
167
+ "grad_norm": 2.8038811683654785,
168
+ "learning_rate": 1.9446676970633695e-05,
169
+ "loss": 3.961,
170
+ "mean_token_accuracy": 0.33416116759181025,
171
+ "num_tokens": 565598.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.05873261205564142,
176
+ "grad_norm": 8.930831909179688,
177
+ "learning_rate": 1.941576506955178e-05,
178
+ "loss": 4.0107,
179
+ "mean_token_accuracy": 0.3261258576065302,
180
+ "num_tokens": 598505.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.061823802163833076,
185
+ "grad_norm": 4.706038475036621,
186
+ "learning_rate": 1.9384853168469862e-05,
187
+ "loss": 3.9231,
188
+ "mean_token_accuracy": 0.33582728281617164,
189
+ "num_tokens": 630284.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.06491499227202473,
194
+ "grad_norm": 2.9916040897369385,
195
+ "learning_rate": 1.9353941267387946e-05,
196
+ "loss": 3.9407,
197
+ "mean_token_accuracy": 0.32844844460487366,
198
+ "num_tokens": 660513.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.06800618238021638,
203
+ "grad_norm": 2.763737678527832,
204
+ "learning_rate": 1.932302936630603e-05,
205
+ "loss": 4.0479,
206
+ "mean_token_accuracy": 0.3248734712600708,
207
+ "num_tokens": 693154.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.07109737248840804,
212
+ "grad_norm": 3.656487464904785,
213
+ "learning_rate": 1.9292117465224113e-05,
214
+ "loss": 3.7843,
215
+ "mean_token_accuracy": 0.34686593189835546,
216
+ "num_tokens": 724686.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.07418856259659969,
221
+ "grad_norm": 2.6140244007110596,
222
+ "learning_rate": 1.9261205564142196e-05,
223
+ "loss": 3.9261,
224
+ "mean_token_accuracy": 0.3310456670820713,
225
+ "num_tokens": 755625.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.07727975270479134,
230
+ "grad_norm": 1.595627784729004,
231
+ "learning_rate": 1.923029366306028e-05,
232
+ "loss": 3.841,
233
+ "mean_token_accuracy": 0.34514380544424056,
234
+ "num_tokens": 785478.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.080370942812983,
239
+ "grad_norm": 2.0485758781433105,
240
+ "learning_rate": 1.9199381761978363e-05,
241
+ "loss": 3.7596,
242
+ "mean_token_accuracy": 0.35417362824082377,
243
+ "num_tokens": 816351.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.08346213292117466,
248
+ "grad_norm": 1.7564281225204468,
249
+ "learning_rate": 1.9168469860896447e-05,
250
+ "loss": 3.7927,
251
+ "mean_token_accuracy": 0.3487237967550755,
252
+ "num_tokens": 849774.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.0865533230293663,
257
+ "grad_norm": 1.5662060976028442,
258
+ "learning_rate": 1.913755795981453e-05,
259
+ "loss": 3.7821,
260
+ "mean_token_accuracy": 0.3511029303073883,
261
+ "num_tokens": 881474.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.08964451313755796,
266
+ "grad_norm": 2.539433002471924,
267
+ "learning_rate": 1.9106646058732614e-05,
268
+ "loss": 3.8613,
269
+ "mean_token_accuracy": 0.3471171148121357,
270
+ "num_tokens": 913459.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.09273570324574962,
275
+ "grad_norm": 4.281046390533447,
276
+ "learning_rate": 1.9075734157650694e-05,
277
+ "loss": 3.8436,
278
+ "mean_token_accuracy": 0.3475985363125801,
279
+ "num_tokens": 947091.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.09582689335394126,
284
+ "grad_norm": 2.379791736602783,
285
+ "learning_rate": 1.904482225656878e-05,
286
+ "loss": 3.8309,
287
+ "mean_token_accuracy": 0.34195478409528735,
288
+ "num_tokens": 983471.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.09891808346213292,
293
+ "grad_norm": 2.4176697731018066,
294
+ "learning_rate": 1.9013910355486865e-05,
295
+ "loss": 3.75,
296
+ "mean_token_accuracy": 0.3633588753640652,
297
+ "num_tokens": 1011575.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.10200927357032458,
302
+ "grad_norm": 3.375523328781128,
303
+ "learning_rate": 1.898299845440495e-05,
304
+ "loss": 3.7556,
305
+ "mean_token_accuracy": 0.36073010191321375,
306
+ "num_tokens": 1042031.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.10510046367851623,
311
+ "grad_norm": 5.099122524261475,
312
+ "learning_rate": 1.895208655332303e-05,
313
+ "loss": 3.6861,
314
+ "mean_token_accuracy": 0.36301063373684883,
315
+ "num_tokens": 1074689.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.10819165378670788,
320
+ "grad_norm": 1.4775930643081665,
321
+ "learning_rate": 1.8921174652241116e-05,
322
+ "loss": 3.6744,
323
+ "mean_token_accuracy": 0.3681299857795238,
324
+ "num_tokens": 1105719.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.11128284389489954,
329
+ "grad_norm": 3.931447744369507,
330
+ "learning_rate": 1.88902627511592e-05,
331
+ "loss": 3.6248,
332
+ "mean_token_accuracy": 0.37062914595007895,
333
+ "num_tokens": 1136185.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.1143740340030912,
338
+ "grad_norm": 2.6153130531311035,
339
+ "learning_rate": 1.885935085007728e-05,
340
+ "loss": 3.6971,
341
+ "mean_token_accuracy": 0.35686987787485125,
342
+ "num_tokens": 1166461.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.11746522411128284,
347
+ "grad_norm": 3.0943849086761475,
348
+ "learning_rate": 1.8828438948995363e-05,
349
+ "loss": 3.7057,
350
+ "mean_token_accuracy": 0.36670113652944564,
351
+ "num_tokens": 1194661.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.1205564142194745,
356
+ "grad_norm": 1.760920524597168,
357
+ "learning_rate": 1.879752704791345e-05,
358
+ "loss": 3.6806,
359
+ "mean_token_accuracy": 0.366318603605032,
360
+ "num_tokens": 1224670.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.12364760432766615,
365
+ "grad_norm": 1.8976062536239624,
366
+ "learning_rate": 1.8766615146831534e-05,
367
+ "loss": 3.7318,
368
+ "mean_token_accuracy": 0.36500929966568946,
369
+ "num_tokens": 1252830.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.1267387944358578,
374
+ "grad_norm": 1.950358510017395,
375
+ "learning_rate": 1.8735703245749614e-05,
376
+ "loss": 3.7106,
377
+ "mean_token_accuracy": 0.3675060346722603,
378
+ "num_tokens": 1285546.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.12982998454404945,
383
+ "grad_norm": 2.707167148590088,
384
+ "learning_rate": 1.8704791344667697e-05,
385
+ "loss": 3.6688,
386
+ "mean_token_accuracy": 0.37156677842140196,
387
+ "num_tokens": 1316466.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.13292117465224113,
392
+ "grad_norm": 2.084510564804077,
393
+ "learning_rate": 1.8673879443585784e-05,
394
+ "loss": 3.6758,
395
+ "mean_token_accuracy": 0.3684880450367928,
396
+ "num_tokens": 1349121.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.13601236476043277,
401
+ "grad_norm": 2.2626636028289795,
402
+ "learning_rate": 1.8642967542503865e-05,
403
+ "loss": 3.7214,
404
+ "mean_token_accuracy": 0.35720510333776473,
405
+ "num_tokens": 1384366.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.1391035548686244,
410
+ "grad_norm": 2.4145290851593018,
411
+ "learning_rate": 1.8612055641421948e-05,
412
+ "loss": 3.7261,
413
+ "mean_token_accuracy": 0.36429562568664553,
414
+ "num_tokens": 1415357.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.14219474497681608,
419
+ "grad_norm": 2.7409212589263916,
420
+ "learning_rate": 1.8581143740340032e-05,
421
+ "loss": 3.7368,
422
+ "mean_token_accuracy": 0.3619597226381302,
423
+ "num_tokens": 1445641.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.14528593508500773,
428
+ "grad_norm": 4.5937275886535645,
429
+ "learning_rate": 1.8550231839258115e-05,
430
+ "loss": 3.6911,
431
+ "mean_token_accuracy": 0.3683665543794632,
432
+ "num_tokens": 1475360.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.14837712519319937,
437
+ "grad_norm": 12.00837516784668,
438
+ "learning_rate": 1.85193199381762e-05,
439
+ "loss": 3.6899,
440
+ "mean_token_accuracy": 0.3648961283266544,
441
+ "num_tokens": 1506048.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.15146831530139104,
446
+ "grad_norm": 2.5790181159973145,
447
+ "learning_rate": 1.8488408037094283e-05,
448
+ "loss": 3.6948,
449
+ "mean_token_accuracy": 0.36360194012522695,
450
+ "num_tokens": 1539998.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.1545595054095827,
455
+ "grad_norm": 1.8515182733535767,
456
+ "learning_rate": 1.8457496136012366e-05,
457
+ "loss": 3.65,
458
+ "mean_token_accuracy": 0.36655392646789553,
459
+ "num_tokens": 1573641.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.15765069551777433,
464
+ "grad_norm": 2.3960251808166504,
465
+ "learning_rate": 1.842658423493045e-05,
466
+ "loss": 3.58,
467
+ "mean_token_accuracy": 0.373273029923439,
468
+ "num_tokens": 1605367.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.160741885625966,
473
+ "grad_norm": 2.578730821609497,
474
+ "learning_rate": 1.8395672333848533e-05,
475
+ "loss": 3.6695,
476
+ "mean_token_accuracy": 0.368588350713253,
477
+ "num_tokens": 1638052.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.16383307573415765,
482
+ "grad_norm": 1.9368691444396973,
483
+ "learning_rate": 1.8364760432766617e-05,
484
+ "loss": 3.6363,
485
+ "mean_token_accuracy": 0.3681382529437542,
486
+ "num_tokens": 1672043.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.16692426584234932,
491
+ "grad_norm": 3.562593698501587,
492
+ "learning_rate": 1.83338485316847e-05,
493
+ "loss": 3.5386,
494
+ "mean_token_accuracy": 0.3801154658198357,
495
+ "num_tokens": 1704817.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.17001545595054096,
500
+ "grad_norm": 1.564929723739624,
501
+ "learning_rate": 1.8302936630602784e-05,
502
+ "loss": 3.6191,
503
+ "mean_token_accuracy": 0.36746986880898475,
504
+ "num_tokens": 1737297.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.1731066460587326,
509
+ "grad_norm": 1.4315626621246338,
510
+ "learning_rate": 1.8272024729520868e-05,
511
+ "loss": 3.592,
512
+ "mean_token_accuracy": 0.3701712526381016,
513
+ "num_tokens": 1770204.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 0.17619783616692428,
518
+ "grad_norm": 6.3183746337890625,
519
+ "learning_rate": 1.824111282843895e-05,
520
+ "loss": 3.6281,
521
+ "mean_token_accuracy": 0.37336429879069327,
522
+ "num_tokens": 1800581.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 0.17928902627511592,
527
+ "grad_norm": 1.3608078956604004,
528
+ "learning_rate": 1.8210200927357035e-05,
529
+ "loss": 3.5644,
530
+ "mean_token_accuracy": 0.37816725075244906,
531
+ "num_tokens": 1834564.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 0.18238021638330756,
536
+ "grad_norm": 11.443370819091797,
537
+ "learning_rate": 1.817928902627512e-05,
538
+ "loss": 3.6134,
539
+ "mean_token_accuracy": 0.37460487633943557,
540
+ "num_tokens": 1866113.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 0.18547140649149924,
545
+ "grad_norm": 2.1869001388549805,
546
+ "learning_rate": 1.8148377125193202e-05,
547
+ "loss": 3.6331,
548
+ "mean_token_accuracy": 0.3722210742533207,
549
+ "num_tokens": 1900627.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 0.18856259659969088,
554
+ "grad_norm": 1.8551387786865234,
555
+ "learning_rate": 1.8117465224111282e-05,
556
+ "loss": 3.5707,
557
+ "mean_token_accuracy": 0.3772866874933243,
558
+ "num_tokens": 1932988.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 0.19165378670788252,
563
+ "grad_norm": 5.528620719909668,
564
+ "learning_rate": 1.808655332302937e-05,
565
+ "loss": 3.6475,
566
+ "mean_token_accuracy": 0.3652547873556614,
567
+ "num_tokens": 1962031.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 0.1947449768160742,
572
+ "grad_norm": 3.566514253616333,
573
+ "learning_rate": 1.8055641421947453e-05,
574
+ "loss": 3.5441,
575
+ "mean_token_accuracy": 0.3836688004434109,
576
+ "num_tokens": 1995070.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 0.19783616692426584,
581
+ "grad_norm": 1.9582892656326294,
582
+ "learning_rate": 1.8024729520865533e-05,
583
+ "loss": 3.5473,
584
+ "mean_token_accuracy": 0.3757788948714733,
585
+ "num_tokens": 2024801.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 0.2009273570324575,
590
+ "grad_norm": 1.7483699321746826,
591
+ "learning_rate": 1.7993817619783616e-05,
592
+ "loss": 3.6133,
593
+ "mean_token_accuracy": 0.37615733668208123,
594
+ "num_tokens": 2054131.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 0.20401854714064915,
599
+ "grad_norm": 1.9855698347091675,
600
+ "learning_rate": 1.7962905718701703e-05,
601
+ "loss": 3.5724,
602
+ "mean_token_accuracy": 0.37552602738142016,
603
+ "num_tokens": 2086218.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 0.2071097372488408,
608
+ "grad_norm": 2.380608558654785,
609
+ "learning_rate": 1.7931993817619787e-05,
610
+ "loss": 3.5476,
611
+ "mean_token_accuracy": 0.37423405200243,
612
+ "num_tokens": 2121389.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 0.21020092735703247,
617
+ "grad_norm": 1.5300630331039429,
618
+ "learning_rate": 1.7901081916537867e-05,
619
+ "loss": 3.668,
620
+ "mean_token_accuracy": 0.3656138554215431,
621
+ "num_tokens": 2152513.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 0.2132921174652241,
626
+ "grad_norm": 1.6176111698150635,
627
+ "learning_rate": 1.787017001545595e-05,
628
+ "loss": 3.6087,
629
+ "mean_token_accuracy": 0.37140627652406694,
630
+ "num_tokens": 2184920.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 0.21638330757341576,
635
+ "grad_norm": 1.401524305343628,
636
+ "learning_rate": 1.7839258114374038e-05,
637
+ "loss": 3.4881,
638
+ "mean_token_accuracy": 0.3919842541217804,
639
+ "num_tokens": 2214174.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 0.21947449768160743,
644
+ "grad_norm": 1.4391794204711914,
645
+ "learning_rate": 1.7808346213292118e-05,
646
+ "loss": 3.5622,
647
+ "mean_token_accuracy": 0.37740132212638855,
648
+ "num_tokens": 2247593.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 0.22256568778979907,
653
+ "grad_norm": 5.364869117736816,
654
+ "learning_rate": 1.77774343122102e-05,
655
+ "loss": 3.5855,
656
+ "mean_token_accuracy": 0.37872459217905996,
657
+ "num_tokens": 2277810.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 0.22565687789799072,
662
+ "grad_norm": 1.3683693408966064,
663
+ "learning_rate": 1.7746522411128285e-05,
664
+ "loss": 3.4653,
665
+ "mean_token_accuracy": 0.38120819330215455,
666
+ "num_tokens": 2309457.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 0.2287480680061824,
671
+ "grad_norm": 1.3637374639511108,
672
+ "learning_rate": 1.771561051004637e-05,
673
+ "loss": 3.5131,
674
+ "mean_token_accuracy": 0.3864184685051441,
675
+ "num_tokens": 2339852.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 0.23183925811437403,
680
+ "grad_norm": 2.7388429641723633,
681
+ "learning_rate": 1.7684698608964452e-05,
682
+ "loss": 3.573,
683
+ "mean_token_accuracy": 0.3795412413775921,
684
+ "num_tokens": 2371444.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 0.23493044822256567,
689
+ "grad_norm": 9.798398971557617,
690
+ "learning_rate": 1.7653786707882536e-05,
691
+ "loss": 3.5441,
692
+ "mean_token_accuracy": 0.38080229982733727,
693
+ "num_tokens": 2401769.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 0.23802163833075735,
698
+ "grad_norm": 2.1416878700256348,
699
+ "learning_rate": 1.762287480680062e-05,
700
+ "loss": 3.5043,
701
+ "mean_token_accuracy": 0.3838089659810066,
702
+ "num_tokens": 2432410.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 0.241112828438949,
707
+ "grad_norm": 2.211545467376709,
708
+ "learning_rate": 1.7591962905718703e-05,
709
+ "loss": 3.5341,
710
+ "mean_token_accuracy": 0.388753118366003,
711
+ "num_tokens": 2462641.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 0.24420401854714066,
716
+ "grad_norm": 5.351387023925781,
717
+ "learning_rate": 1.7561051004636787e-05,
718
+ "loss": 3.5031,
719
+ "mean_token_accuracy": 0.3821559719741344,
720
+ "num_tokens": 2496459.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 0.2472952086553323,
725
+ "grad_norm": 2.3877508640289307,
726
+ "learning_rate": 1.753013910355487e-05,
727
+ "loss": 3.5589,
728
+ "mean_token_accuracy": 0.3789012677967548,
729
+ "num_tokens": 2526708.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 0.250386398763524,
734
+ "grad_norm": 1.2962738275527954,
735
+ "learning_rate": 1.7499227202472954e-05,
736
+ "loss": 3.505,
737
+ "mean_token_accuracy": 0.3823570780456066,
738
+ "num_tokens": 2561229.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 0.2534775888717156,
743
+ "grad_norm": 1.8147461414337158,
744
+ "learning_rate": 1.7468315301391037e-05,
745
+ "loss": 3.4861,
746
+ "mean_token_accuracy": 0.3802278622984886,
747
+ "num_tokens": 2592914.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 0.25656877897990726,
752
+ "grad_norm": 3.4690439701080322,
753
+ "learning_rate": 1.743740340030912e-05,
754
+ "loss": 3.5202,
755
+ "mean_token_accuracy": 0.38112854287028314,
756
+ "num_tokens": 2627185.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 0.2596599690880989,
761
+ "grad_norm": 4.353795051574707,
762
+ "learning_rate": 1.7406491499227205e-05,
763
+ "loss": 3.5667,
764
+ "mean_token_accuracy": 0.37200469225645066,
765
+ "num_tokens": 2660839.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 0.26275115919629055,
770
+ "grad_norm": 2.6319100856781006,
771
+ "learning_rate": 1.7375579598145288e-05,
772
+ "loss": 3.464,
773
+ "mean_token_accuracy": 0.38627258986234664,
774
+ "num_tokens": 2690042.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 0.26584234930448225,
779
+ "grad_norm": 1.6525287628173828,
780
+ "learning_rate": 1.7344667697063372e-05,
781
+ "loss": 3.5616,
782
+ "mean_token_accuracy": 0.3763453342020512,
783
+ "num_tokens": 2724354.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 0.2689335394126739,
788
+ "grad_norm": 1.9737221002578735,
789
+ "learning_rate": 1.7313755795981455e-05,
790
+ "loss": 3.5147,
791
+ "mean_token_accuracy": 0.38340551406145096,
792
+ "num_tokens": 2755591.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 0.27202472952086554,
797
+ "grad_norm": 2.5324320793151855,
798
+ "learning_rate": 1.7282843894899536e-05,
799
+ "loss": 3.4227,
800
+ "mean_token_accuracy": 0.3909419260919094,
801
+ "num_tokens": 2786566.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 0.2751159196290572,
806
+ "grad_norm": 3.3844659328460693,
807
+ "learning_rate": 1.7251931993817623e-05,
808
+ "loss": 3.5561,
809
+ "mean_token_accuracy": 0.37579271346330645,
810
+ "num_tokens": 2814171.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 0.2782071097372488,
815
+ "grad_norm": 2.7620153427124023,
816
+ "learning_rate": 1.7221020092735706e-05,
817
+ "loss": 3.5042,
818
+ "mean_token_accuracy": 0.3822973191738129,
819
+ "num_tokens": 2844402.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 0.28129829984544047,
824
+ "grad_norm": 2.089118719100952,
825
+ "learning_rate": 1.7190108191653786e-05,
826
+ "loss": 3.4583,
827
+ "mean_token_accuracy": 0.39270223304629326,
828
+ "num_tokens": 2873728.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 0.28438948995363217,
833
+ "grad_norm": 2.496480941772461,
834
+ "learning_rate": 1.715919629057187e-05,
835
+ "loss": 3.4466,
836
+ "mean_token_accuracy": 0.3897694177925587,
837
+ "num_tokens": 2905000.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 0.2874806800618238,
842
+ "grad_norm": 2.162785053253174,
843
+ "learning_rate": 1.7128284389489957e-05,
844
+ "loss": 3.4075,
845
+ "mean_token_accuracy": 0.39374888986349105,
846
+ "num_tokens": 2935706.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 0.29057187017001546,
851
+ "grad_norm": 1.5756185054779053,
852
+ "learning_rate": 1.709737248840804e-05,
853
+ "loss": 3.4396,
854
+ "mean_token_accuracy": 0.38780966177582743,
855
+ "num_tokens": 2966477.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 0.2936630602782071,
860
+ "grad_norm": 5.708625316619873,
861
+ "learning_rate": 1.706646058732612e-05,
862
+ "loss": 3.4185,
863
+ "mean_token_accuracy": 0.39326486811041833,
864
+ "num_tokens": 2996439.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 0.29675425038639874,
869
+ "grad_norm": 2.8689119815826416,
870
+ "learning_rate": 1.7035548686244204e-05,
871
+ "loss": 3.468,
872
+ "mean_token_accuracy": 0.3818407289683819,
873
+ "num_tokens": 3026332.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 0.29984544049459044,
878
+ "grad_norm": 4.020405292510986,
879
+ "learning_rate": 1.700463678516229e-05,
880
+ "loss": 3.4943,
881
+ "mean_token_accuracy": 0.3888055384159088,
882
+ "num_tokens": 3057606.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 0.3029366306027821,
887
+ "grad_norm": 1.8991873264312744,
888
+ "learning_rate": 1.697372488408037e-05,
889
+ "loss": 3.4939,
890
+ "mean_token_accuracy": 0.38414665684103966,
891
+ "num_tokens": 3089449.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 0.30602782071097373,
896
+ "grad_norm": 13.78397274017334,
897
+ "learning_rate": 1.6942812982998455e-05,
898
+ "loss": 3.5487,
899
+ "mean_token_accuracy": 0.3769320294260979,
900
+ "num_tokens": 3119537.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 0.3091190108191654,
905
+ "grad_norm": 1.599820613861084,
906
+ "learning_rate": 1.691190108191654e-05,
907
+ "loss": 3.465,
908
+ "mean_token_accuracy": 0.3884730890393257,
909
+ "num_tokens": 3151905.0,
910
+ "step": 1000
911
+ }
912
+ ],
913
+ "logging_steps": 10,
914
+ "max_steps": 6470,
915
+ "num_input_tokens_seen": 0,
916
+ "num_train_epochs": 2,
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": 1821690846111744.0,
931
+ "train_batch_size": 4,
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:f7b555e5b9b497f109f303fcefbaa4f3565d3b5b6b83ddc274743aea488d2cd5
3
+ size 5624
checkpoint-1500/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": 256,
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-1500/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-1500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57a037b77e736d0cbce3de584e974ede33f1876d2e22c472a84160da11d6db2c
3
+ size 309900448
checkpoint-1500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d630059534c1c4aca4b9f256d027c3bffa52109f7777adbe0601cce854491bd8
3
+ size 619836730
checkpoint-1500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
checkpoint-1500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f83f3014faca7ba00ad3919eecbe91a54ae3b1b9b6e9661b24537f0d509b43cb
3
+ size 1064
checkpoint-1500/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-1500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1500/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-1500/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-1500/trainer_state.json ADDED
@@ -0,0 +1,1384 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.46367851622874806,
6
+ "eval_steps": 500,
7
+ "global_step": 1500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0030911901081916537,
14
+ "grad_norm": 3.058769941329956,
15
+ "learning_rate": 1.9972179289026277e-05,
16
+ "loss": 5.0209,
17
+ "mean_token_accuracy": 0.21833103336393833,
18
+ "num_tokens": 28941.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.0061823802163833074,
23
+ "grad_norm": 2.485980749130249,
24
+ "learning_rate": 1.994126738794436e-05,
25
+ "loss": 4.7631,
26
+ "mean_token_accuracy": 0.23807235918939113,
27
+ "num_tokens": 60990.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.00927357032457496,
32
+ "grad_norm": 5.288544178009033,
33
+ "learning_rate": 1.9910355486862444e-05,
34
+ "loss": 4.6799,
35
+ "mean_token_accuracy": 0.24858475103974342,
36
+ "num_tokens": 90602.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.012364760432766615,
41
+ "grad_norm": 3.226719379425049,
42
+ "learning_rate": 1.9879443585780528e-05,
43
+ "loss": 4.5883,
44
+ "mean_token_accuracy": 0.2555678006261587,
45
+ "num_tokens": 121132.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.015455950540958269,
50
+ "grad_norm": 2.027597188949585,
51
+ "learning_rate": 1.984853168469861e-05,
52
+ "loss": 4.5125,
53
+ "mean_token_accuracy": 0.26637452803552153,
54
+ "num_tokens": 154244.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.01854714064914992,
59
+ "grad_norm": 1.6715962886810303,
60
+ "learning_rate": 1.9817619783616695e-05,
61
+ "loss": 4.4224,
62
+ "mean_token_accuracy": 0.27198897041380404,
63
+ "num_tokens": 185618.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.021638330757341576,
68
+ "grad_norm": 3.0472540855407715,
69
+ "learning_rate": 1.9786707882534775e-05,
70
+ "loss": 4.3464,
71
+ "mean_token_accuracy": 0.2801041007041931,
72
+ "num_tokens": 220138.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.02472952086553323,
77
+ "grad_norm": 2.74045729637146,
78
+ "learning_rate": 1.9755795981452862e-05,
79
+ "loss": 4.4569,
80
+ "mean_token_accuracy": 0.2733523309230804,
81
+ "num_tokens": 252919.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.027820710973724884,
86
+ "grad_norm": 1.9274535179138184,
87
+ "learning_rate": 1.9724884080370946e-05,
88
+ "loss": 4.2569,
89
+ "mean_token_accuracy": 0.2858310595154762,
90
+ "num_tokens": 284739.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.030911901081916538,
95
+ "grad_norm": 1.906964898109436,
96
+ "learning_rate": 1.9693972179289026e-05,
97
+ "loss": 4.2144,
98
+ "mean_token_accuracy": 0.2953581381589174,
99
+ "num_tokens": 315186.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.03400309119010819,
104
+ "grad_norm": 2.042823314666748,
105
+ "learning_rate": 1.966306027820711e-05,
106
+ "loss": 4.2595,
107
+ "mean_token_accuracy": 0.287172843888402,
108
+ "num_tokens": 347507.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.03709428129829984,
113
+ "grad_norm": 1.9889057874679565,
114
+ "learning_rate": 1.9632148377125197e-05,
115
+ "loss": 4.1643,
116
+ "mean_token_accuracy": 0.30306958928704264,
117
+ "num_tokens": 380331.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.0401854714064915,
122
+ "grad_norm": 1.7952933311462402,
123
+ "learning_rate": 1.9601236476043277e-05,
124
+ "loss": 4.1334,
125
+ "mean_token_accuracy": 0.3080880597233772,
126
+ "num_tokens": 410947.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.04327666151468315,
131
+ "grad_norm": 3.8508658409118652,
132
+ "learning_rate": 1.957032457496136e-05,
133
+ "loss": 4.0262,
134
+ "mean_token_accuracy": 0.3118089348077774,
135
+ "num_tokens": 444130.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.04636785162287481,
140
+ "grad_norm": 31.280546188354492,
141
+ "learning_rate": 1.9539412673879444e-05,
142
+ "loss": 3.9889,
143
+ "mean_token_accuracy": 0.32304659858345985,
144
+ "num_tokens": 474570.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.04945904173106646,
149
+ "grad_norm": 2.1569416522979736,
150
+ "learning_rate": 1.950850077279753e-05,
151
+ "loss": 4.1053,
152
+ "mean_token_accuracy": 0.31258094161748884,
153
+ "num_tokens": 505047.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.05255023183925812,
158
+ "grad_norm": 1.808377981185913,
159
+ "learning_rate": 1.947758887171561e-05,
160
+ "loss": 3.9711,
161
+ "mean_token_accuracy": 0.33312112018465995,
162
+ "num_tokens": 533187.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.05564142194744977,
167
+ "grad_norm": 2.8038811683654785,
168
+ "learning_rate": 1.9446676970633695e-05,
169
+ "loss": 3.961,
170
+ "mean_token_accuracy": 0.33416116759181025,
171
+ "num_tokens": 565598.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.05873261205564142,
176
+ "grad_norm": 8.930831909179688,
177
+ "learning_rate": 1.941576506955178e-05,
178
+ "loss": 4.0107,
179
+ "mean_token_accuracy": 0.3261258576065302,
180
+ "num_tokens": 598505.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.061823802163833076,
185
+ "grad_norm": 4.706038475036621,
186
+ "learning_rate": 1.9384853168469862e-05,
187
+ "loss": 3.9231,
188
+ "mean_token_accuracy": 0.33582728281617164,
189
+ "num_tokens": 630284.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.06491499227202473,
194
+ "grad_norm": 2.9916040897369385,
195
+ "learning_rate": 1.9353941267387946e-05,
196
+ "loss": 3.9407,
197
+ "mean_token_accuracy": 0.32844844460487366,
198
+ "num_tokens": 660513.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.06800618238021638,
203
+ "grad_norm": 2.763737678527832,
204
+ "learning_rate": 1.932302936630603e-05,
205
+ "loss": 4.0479,
206
+ "mean_token_accuracy": 0.3248734712600708,
207
+ "num_tokens": 693154.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.07109737248840804,
212
+ "grad_norm": 3.656487464904785,
213
+ "learning_rate": 1.9292117465224113e-05,
214
+ "loss": 3.7843,
215
+ "mean_token_accuracy": 0.34686593189835546,
216
+ "num_tokens": 724686.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.07418856259659969,
221
+ "grad_norm": 2.6140244007110596,
222
+ "learning_rate": 1.9261205564142196e-05,
223
+ "loss": 3.9261,
224
+ "mean_token_accuracy": 0.3310456670820713,
225
+ "num_tokens": 755625.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.07727975270479134,
230
+ "grad_norm": 1.595627784729004,
231
+ "learning_rate": 1.923029366306028e-05,
232
+ "loss": 3.841,
233
+ "mean_token_accuracy": 0.34514380544424056,
234
+ "num_tokens": 785478.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.080370942812983,
239
+ "grad_norm": 2.0485758781433105,
240
+ "learning_rate": 1.9199381761978363e-05,
241
+ "loss": 3.7596,
242
+ "mean_token_accuracy": 0.35417362824082377,
243
+ "num_tokens": 816351.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.08346213292117466,
248
+ "grad_norm": 1.7564281225204468,
249
+ "learning_rate": 1.9168469860896447e-05,
250
+ "loss": 3.7927,
251
+ "mean_token_accuracy": 0.3487237967550755,
252
+ "num_tokens": 849774.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.0865533230293663,
257
+ "grad_norm": 1.5662060976028442,
258
+ "learning_rate": 1.913755795981453e-05,
259
+ "loss": 3.7821,
260
+ "mean_token_accuracy": 0.3511029303073883,
261
+ "num_tokens": 881474.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.08964451313755796,
266
+ "grad_norm": 2.539433002471924,
267
+ "learning_rate": 1.9106646058732614e-05,
268
+ "loss": 3.8613,
269
+ "mean_token_accuracy": 0.3471171148121357,
270
+ "num_tokens": 913459.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.09273570324574962,
275
+ "grad_norm": 4.281046390533447,
276
+ "learning_rate": 1.9075734157650694e-05,
277
+ "loss": 3.8436,
278
+ "mean_token_accuracy": 0.3475985363125801,
279
+ "num_tokens": 947091.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.09582689335394126,
284
+ "grad_norm": 2.379791736602783,
285
+ "learning_rate": 1.904482225656878e-05,
286
+ "loss": 3.8309,
287
+ "mean_token_accuracy": 0.34195478409528735,
288
+ "num_tokens": 983471.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.09891808346213292,
293
+ "grad_norm": 2.4176697731018066,
294
+ "learning_rate": 1.9013910355486865e-05,
295
+ "loss": 3.75,
296
+ "mean_token_accuracy": 0.3633588753640652,
297
+ "num_tokens": 1011575.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.10200927357032458,
302
+ "grad_norm": 3.375523328781128,
303
+ "learning_rate": 1.898299845440495e-05,
304
+ "loss": 3.7556,
305
+ "mean_token_accuracy": 0.36073010191321375,
306
+ "num_tokens": 1042031.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.10510046367851623,
311
+ "grad_norm": 5.099122524261475,
312
+ "learning_rate": 1.895208655332303e-05,
313
+ "loss": 3.6861,
314
+ "mean_token_accuracy": 0.36301063373684883,
315
+ "num_tokens": 1074689.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.10819165378670788,
320
+ "grad_norm": 1.4775930643081665,
321
+ "learning_rate": 1.8921174652241116e-05,
322
+ "loss": 3.6744,
323
+ "mean_token_accuracy": 0.3681299857795238,
324
+ "num_tokens": 1105719.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.11128284389489954,
329
+ "grad_norm": 3.931447744369507,
330
+ "learning_rate": 1.88902627511592e-05,
331
+ "loss": 3.6248,
332
+ "mean_token_accuracy": 0.37062914595007895,
333
+ "num_tokens": 1136185.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.1143740340030912,
338
+ "grad_norm": 2.6153130531311035,
339
+ "learning_rate": 1.885935085007728e-05,
340
+ "loss": 3.6971,
341
+ "mean_token_accuracy": 0.35686987787485125,
342
+ "num_tokens": 1166461.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.11746522411128284,
347
+ "grad_norm": 3.0943849086761475,
348
+ "learning_rate": 1.8828438948995363e-05,
349
+ "loss": 3.7057,
350
+ "mean_token_accuracy": 0.36670113652944564,
351
+ "num_tokens": 1194661.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.1205564142194745,
356
+ "grad_norm": 1.760920524597168,
357
+ "learning_rate": 1.879752704791345e-05,
358
+ "loss": 3.6806,
359
+ "mean_token_accuracy": 0.366318603605032,
360
+ "num_tokens": 1224670.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.12364760432766615,
365
+ "grad_norm": 1.8976062536239624,
366
+ "learning_rate": 1.8766615146831534e-05,
367
+ "loss": 3.7318,
368
+ "mean_token_accuracy": 0.36500929966568946,
369
+ "num_tokens": 1252830.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.1267387944358578,
374
+ "grad_norm": 1.950358510017395,
375
+ "learning_rate": 1.8735703245749614e-05,
376
+ "loss": 3.7106,
377
+ "mean_token_accuracy": 0.3675060346722603,
378
+ "num_tokens": 1285546.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.12982998454404945,
383
+ "grad_norm": 2.707167148590088,
384
+ "learning_rate": 1.8704791344667697e-05,
385
+ "loss": 3.6688,
386
+ "mean_token_accuracy": 0.37156677842140196,
387
+ "num_tokens": 1316466.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.13292117465224113,
392
+ "grad_norm": 2.084510564804077,
393
+ "learning_rate": 1.8673879443585784e-05,
394
+ "loss": 3.6758,
395
+ "mean_token_accuracy": 0.3684880450367928,
396
+ "num_tokens": 1349121.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.13601236476043277,
401
+ "grad_norm": 2.2626636028289795,
402
+ "learning_rate": 1.8642967542503865e-05,
403
+ "loss": 3.7214,
404
+ "mean_token_accuracy": 0.35720510333776473,
405
+ "num_tokens": 1384366.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.1391035548686244,
410
+ "grad_norm": 2.4145290851593018,
411
+ "learning_rate": 1.8612055641421948e-05,
412
+ "loss": 3.7261,
413
+ "mean_token_accuracy": 0.36429562568664553,
414
+ "num_tokens": 1415357.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.14219474497681608,
419
+ "grad_norm": 2.7409212589263916,
420
+ "learning_rate": 1.8581143740340032e-05,
421
+ "loss": 3.7368,
422
+ "mean_token_accuracy": 0.3619597226381302,
423
+ "num_tokens": 1445641.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.14528593508500773,
428
+ "grad_norm": 4.5937275886535645,
429
+ "learning_rate": 1.8550231839258115e-05,
430
+ "loss": 3.6911,
431
+ "mean_token_accuracy": 0.3683665543794632,
432
+ "num_tokens": 1475360.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.14837712519319937,
437
+ "grad_norm": 12.00837516784668,
438
+ "learning_rate": 1.85193199381762e-05,
439
+ "loss": 3.6899,
440
+ "mean_token_accuracy": 0.3648961283266544,
441
+ "num_tokens": 1506048.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.15146831530139104,
446
+ "grad_norm": 2.5790181159973145,
447
+ "learning_rate": 1.8488408037094283e-05,
448
+ "loss": 3.6948,
449
+ "mean_token_accuracy": 0.36360194012522695,
450
+ "num_tokens": 1539998.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.1545595054095827,
455
+ "grad_norm": 1.8515182733535767,
456
+ "learning_rate": 1.8457496136012366e-05,
457
+ "loss": 3.65,
458
+ "mean_token_accuracy": 0.36655392646789553,
459
+ "num_tokens": 1573641.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.15765069551777433,
464
+ "grad_norm": 2.3960251808166504,
465
+ "learning_rate": 1.842658423493045e-05,
466
+ "loss": 3.58,
467
+ "mean_token_accuracy": 0.373273029923439,
468
+ "num_tokens": 1605367.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.160741885625966,
473
+ "grad_norm": 2.578730821609497,
474
+ "learning_rate": 1.8395672333848533e-05,
475
+ "loss": 3.6695,
476
+ "mean_token_accuracy": 0.368588350713253,
477
+ "num_tokens": 1638052.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.16383307573415765,
482
+ "grad_norm": 1.9368691444396973,
483
+ "learning_rate": 1.8364760432766617e-05,
484
+ "loss": 3.6363,
485
+ "mean_token_accuracy": 0.3681382529437542,
486
+ "num_tokens": 1672043.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.16692426584234932,
491
+ "grad_norm": 3.562593698501587,
492
+ "learning_rate": 1.83338485316847e-05,
493
+ "loss": 3.5386,
494
+ "mean_token_accuracy": 0.3801154658198357,
495
+ "num_tokens": 1704817.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.17001545595054096,
500
+ "grad_norm": 1.564929723739624,
501
+ "learning_rate": 1.8302936630602784e-05,
502
+ "loss": 3.6191,
503
+ "mean_token_accuracy": 0.36746986880898475,
504
+ "num_tokens": 1737297.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.1731066460587326,
509
+ "grad_norm": 1.4315626621246338,
510
+ "learning_rate": 1.8272024729520868e-05,
511
+ "loss": 3.592,
512
+ "mean_token_accuracy": 0.3701712526381016,
513
+ "num_tokens": 1770204.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 0.17619783616692428,
518
+ "grad_norm": 6.3183746337890625,
519
+ "learning_rate": 1.824111282843895e-05,
520
+ "loss": 3.6281,
521
+ "mean_token_accuracy": 0.37336429879069327,
522
+ "num_tokens": 1800581.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 0.17928902627511592,
527
+ "grad_norm": 1.3608078956604004,
528
+ "learning_rate": 1.8210200927357035e-05,
529
+ "loss": 3.5644,
530
+ "mean_token_accuracy": 0.37816725075244906,
531
+ "num_tokens": 1834564.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 0.18238021638330756,
536
+ "grad_norm": 11.443370819091797,
537
+ "learning_rate": 1.817928902627512e-05,
538
+ "loss": 3.6134,
539
+ "mean_token_accuracy": 0.37460487633943557,
540
+ "num_tokens": 1866113.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 0.18547140649149924,
545
+ "grad_norm": 2.1869001388549805,
546
+ "learning_rate": 1.8148377125193202e-05,
547
+ "loss": 3.6331,
548
+ "mean_token_accuracy": 0.3722210742533207,
549
+ "num_tokens": 1900627.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 0.18856259659969088,
554
+ "grad_norm": 1.8551387786865234,
555
+ "learning_rate": 1.8117465224111282e-05,
556
+ "loss": 3.5707,
557
+ "mean_token_accuracy": 0.3772866874933243,
558
+ "num_tokens": 1932988.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 0.19165378670788252,
563
+ "grad_norm": 5.528620719909668,
564
+ "learning_rate": 1.808655332302937e-05,
565
+ "loss": 3.6475,
566
+ "mean_token_accuracy": 0.3652547873556614,
567
+ "num_tokens": 1962031.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 0.1947449768160742,
572
+ "grad_norm": 3.566514253616333,
573
+ "learning_rate": 1.8055641421947453e-05,
574
+ "loss": 3.5441,
575
+ "mean_token_accuracy": 0.3836688004434109,
576
+ "num_tokens": 1995070.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 0.19783616692426584,
581
+ "grad_norm": 1.9582892656326294,
582
+ "learning_rate": 1.8024729520865533e-05,
583
+ "loss": 3.5473,
584
+ "mean_token_accuracy": 0.3757788948714733,
585
+ "num_tokens": 2024801.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 0.2009273570324575,
590
+ "grad_norm": 1.7483699321746826,
591
+ "learning_rate": 1.7993817619783616e-05,
592
+ "loss": 3.6133,
593
+ "mean_token_accuracy": 0.37615733668208123,
594
+ "num_tokens": 2054131.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 0.20401854714064915,
599
+ "grad_norm": 1.9855698347091675,
600
+ "learning_rate": 1.7962905718701703e-05,
601
+ "loss": 3.5724,
602
+ "mean_token_accuracy": 0.37552602738142016,
603
+ "num_tokens": 2086218.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 0.2071097372488408,
608
+ "grad_norm": 2.380608558654785,
609
+ "learning_rate": 1.7931993817619787e-05,
610
+ "loss": 3.5476,
611
+ "mean_token_accuracy": 0.37423405200243,
612
+ "num_tokens": 2121389.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 0.21020092735703247,
617
+ "grad_norm": 1.5300630331039429,
618
+ "learning_rate": 1.7901081916537867e-05,
619
+ "loss": 3.668,
620
+ "mean_token_accuracy": 0.3656138554215431,
621
+ "num_tokens": 2152513.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 0.2132921174652241,
626
+ "grad_norm": 1.6176111698150635,
627
+ "learning_rate": 1.787017001545595e-05,
628
+ "loss": 3.6087,
629
+ "mean_token_accuracy": 0.37140627652406694,
630
+ "num_tokens": 2184920.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 0.21638330757341576,
635
+ "grad_norm": 1.401524305343628,
636
+ "learning_rate": 1.7839258114374038e-05,
637
+ "loss": 3.4881,
638
+ "mean_token_accuracy": 0.3919842541217804,
639
+ "num_tokens": 2214174.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 0.21947449768160743,
644
+ "grad_norm": 1.4391794204711914,
645
+ "learning_rate": 1.7808346213292118e-05,
646
+ "loss": 3.5622,
647
+ "mean_token_accuracy": 0.37740132212638855,
648
+ "num_tokens": 2247593.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 0.22256568778979907,
653
+ "grad_norm": 5.364869117736816,
654
+ "learning_rate": 1.77774343122102e-05,
655
+ "loss": 3.5855,
656
+ "mean_token_accuracy": 0.37872459217905996,
657
+ "num_tokens": 2277810.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 0.22565687789799072,
662
+ "grad_norm": 1.3683693408966064,
663
+ "learning_rate": 1.7746522411128285e-05,
664
+ "loss": 3.4653,
665
+ "mean_token_accuracy": 0.38120819330215455,
666
+ "num_tokens": 2309457.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 0.2287480680061824,
671
+ "grad_norm": 1.3637374639511108,
672
+ "learning_rate": 1.771561051004637e-05,
673
+ "loss": 3.5131,
674
+ "mean_token_accuracy": 0.3864184685051441,
675
+ "num_tokens": 2339852.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 0.23183925811437403,
680
+ "grad_norm": 2.7388429641723633,
681
+ "learning_rate": 1.7684698608964452e-05,
682
+ "loss": 3.573,
683
+ "mean_token_accuracy": 0.3795412413775921,
684
+ "num_tokens": 2371444.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 0.23493044822256567,
689
+ "grad_norm": 9.798398971557617,
690
+ "learning_rate": 1.7653786707882536e-05,
691
+ "loss": 3.5441,
692
+ "mean_token_accuracy": 0.38080229982733727,
693
+ "num_tokens": 2401769.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 0.23802163833075735,
698
+ "grad_norm": 2.1416878700256348,
699
+ "learning_rate": 1.762287480680062e-05,
700
+ "loss": 3.5043,
701
+ "mean_token_accuracy": 0.3838089659810066,
702
+ "num_tokens": 2432410.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 0.241112828438949,
707
+ "grad_norm": 2.211545467376709,
708
+ "learning_rate": 1.7591962905718703e-05,
709
+ "loss": 3.5341,
710
+ "mean_token_accuracy": 0.388753118366003,
711
+ "num_tokens": 2462641.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 0.24420401854714066,
716
+ "grad_norm": 5.351387023925781,
717
+ "learning_rate": 1.7561051004636787e-05,
718
+ "loss": 3.5031,
719
+ "mean_token_accuracy": 0.3821559719741344,
720
+ "num_tokens": 2496459.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 0.2472952086553323,
725
+ "grad_norm": 2.3877508640289307,
726
+ "learning_rate": 1.753013910355487e-05,
727
+ "loss": 3.5589,
728
+ "mean_token_accuracy": 0.3789012677967548,
729
+ "num_tokens": 2526708.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 0.250386398763524,
734
+ "grad_norm": 1.2962738275527954,
735
+ "learning_rate": 1.7499227202472954e-05,
736
+ "loss": 3.505,
737
+ "mean_token_accuracy": 0.3823570780456066,
738
+ "num_tokens": 2561229.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 0.2534775888717156,
743
+ "grad_norm": 1.8147461414337158,
744
+ "learning_rate": 1.7468315301391037e-05,
745
+ "loss": 3.4861,
746
+ "mean_token_accuracy": 0.3802278622984886,
747
+ "num_tokens": 2592914.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 0.25656877897990726,
752
+ "grad_norm": 3.4690439701080322,
753
+ "learning_rate": 1.743740340030912e-05,
754
+ "loss": 3.5202,
755
+ "mean_token_accuracy": 0.38112854287028314,
756
+ "num_tokens": 2627185.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 0.2596599690880989,
761
+ "grad_norm": 4.353795051574707,
762
+ "learning_rate": 1.7406491499227205e-05,
763
+ "loss": 3.5667,
764
+ "mean_token_accuracy": 0.37200469225645066,
765
+ "num_tokens": 2660839.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 0.26275115919629055,
770
+ "grad_norm": 2.6319100856781006,
771
+ "learning_rate": 1.7375579598145288e-05,
772
+ "loss": 3.464,
773
+ "mean_token_accuracy": 0.38627258986234664,
774
+ "num_tokens": 2690042.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 0.26584234930448225,
779
+ "grad_norm": 1.6525287628173828,
780
+ "learning_rate": 1.7344667697063372e-05,
781
+ "loss": 3.5616,
782
+ "mean_token_accuracy": 0.3763453342020512,
783
+ "num_tokens": 2724354.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 0.2689335394126739,
788
+ "grad_norm": 1.9737221002578735,
789
+ "learning_rate": 1.7313755795981455e-05,
790
+ "loss": 3.5147,
791
+ "mean_token_accuracy": 0.38340551406145096,
792
+ "num_tokens": 2755591.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 0.27202472952086554,
797
+ "grad_norm": 2.5324320793151855,
798
+ "learning_rate": 1.7282843894899536e-05,
799
+ "loss": 3.4227,
800
+ "mean_token_accuracy": 0.3909419260919094,
801
+ "num_tokens": 2786566.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 0.2751159196290572,
806
+ "grad_norm": 3.3844659328460693,
807
+ "learning_rate": 1.7251931993817623e-05,
808
+ "loss": 3.5561,
809
+ "mean_token_accuracy": 0.37579271346330645,
810
+ "num_tokens": 2814171.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 0.2782071097372488,
815
+ "grad_norm": 2.7620153427124023,
816
+ "learning_rate": 1.7221020092735706e-05,
817
+ "loss": 3.5042,
818
+ "mean_token_accuracy": 0.3822973191738129,
819
+ "num_tokens": 2844402.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 0.28129829984544047,
824
+ "grad_norm": 2.089118719100952,
825
+ "learning_rate": 1.7190108191653786e-05,
826
+ "loss": 3.4583,
827
+ "mean_token_accuracy": 0.39270223304629326,
828
+ "num_tokens": 2873728.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 0.28438948995363217,
833
+ "grad_norm": 2.496480941772461,
834
+ "learning_rate": 1.715919629057187e-05,
835
+ "loss": 3.4466,
836
+ "mean_token_accuracy": 0.3897694177925587,
837
+ "num_tokens": 2905000.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 0.2874806800618238,
842
+ "grad_norm": 2.162785053253174,
843
+ "learning_rate": 1.7128284389489957e-05,
844
+ "loss": 3.4075,
845
+ "mean_token_accuracy": 0.39374888986349105,
846
+ "num_tokens": 2935706.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 0.29057187017001546,
851
+ "grad_norm": 1.5756185054779053,
852
+ "learning_rate": 1.709737248840804e-05,
853
+ "loss": 3.4396,
854
+ "mean_token_accuracy": 0.38780966177582743,
855
+ "num_tokens": 2966477.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 0.2936630602782071,
860
+ "grad_norm": 5.708625316619873,
861
+ "learning_rate": 1.706646058732612e-05,
862
+ "loss": 3.4185,
863
+ "mean_token_accuracy": 0.39326486811041833,
864
+ "num_tokens": 2996439.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 0.29675425038639874,
869
+ "grad_norm": 2.8689119815826416,
870
+ "learning_rate": 1.7035548686244204e-05,
871
+ "loss": 3.468,
872
+ "mean_token_accuracy": 0.3818407289683819,
873
+ "num_tokens": 3026332.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 0.29984544049459044,
878
+ "grad_norm": 4.020405292510986,
879
+ "learning_rate": 1.700463678516229e-05,
880
+ "loss": 3.4943,
881
+ "mean_token_accuracy": 0.3888055384159088,
882
+ "num_tokens": 3057606.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 0.3029366306027821,
887
+ "grad_norm": 1.8991873264312744,
888
+ "learning_rate": 1.697372488408037e-05,
889
+ "loss": 3.4939,
890
+ "mean_token_accuracy": 0.38414665684103966,
891
+ "num_tokens": 3089449.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 0.30602782071097373,
896
+ "grad_norm": 13.78397274017334,
897
+ "learning_rate": 1.6942812982998455e-05,
898
+ "loss": 3.5487,
899
+ "mean_token_accuracy": 0.3769320294260979,
900
+ "num_tokens": 3119537.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 0.3091190108191654,
905
+ "grad_norm": 1.599820613861084,
906
+ "learning_rate": 1.691190108191654e-05,
907
+ "loss": 3.465,
908
+ "mean_token_accuracy": 0.3884730890393257,
909
+ "num_tokens": 3151905.0,
910
+ "step": 1000
911
+ },
912
+ {
913
+ "epoch": 0.312210200927357,
914
+ "grad_norm": 6.015178680419922,
915
+ "learning_rate": 1.6880989180834622e-05,
916
+ "loss": 3.5043,
917
+ "mean_token_accuracy": 0.38035417571663854,
918
+ "num_tokens": 3184277.0,
919
+ "step": 1010
920
+ },
921
+ {
922
+ "epoch": 0.31530139103554866,
923
+ "grad_norm": 1.854093313217163,
924
+ "learning_rate": 1.6850077279752706e-05,
925
+ "loss": 3.4233,
926
+ "mean_token_accuracy": 0.389019088447094,
927
+ "num_tokens": 3217455.0,
928
+ "step": 1020
929
+ },
930
+ {
931
+ "epoch": 0.31839258114374036,
932
+ "grad_norm": 3.024531126022339,
933
+ "learning_rate": 1.681916537867079e-05,
934
+ "loss": 3.4768,
935
+ "mean_token_accuracy": 0.39151332527399063,
936
+ "num_tokens": 3251326.0,
937
+ "step": 1030
938
+ },
939
+ {
940
+ "epoch": 0.321483771251932,
941
+ "grad_norm": 1.8012628555297852,
942
+ "learning_rate": 1.6788253477588873e-05,
943
+ "loss": 3.4955,
944
+ "mean_token_accuracy": 0.3855120025575161,
945
+ "num_tokens": 3283488.0,
946
+ "step": 1040
947
+ },
948
+ {
949
+ "epoch": 0.32457496136012365,
950
+ "grad_norm": 1.4446407556533813,
951
+ "learning_rate": 1.6757341576506957e-05,
952
+ "loss": 3.4023,
953
+ "mean_token_accuracy": 0.3883494645357132,
954
+ "num_tokens": 3315970.0,
955
+ "step": 1050
956
+ },
957
+ {
958
+ "epoch": 0.3276661514683153,
959
+ "grad_norm": 2.132194995880127,
960
+ "learning_rate": 1.672642967542504e-05,
961
+ "loss": 3.5195,
962
+ "mean_token_accuracy": 0.38440208286046984,
963
+ "num_tokens": 3348645.0,
964
+ "step": 1060
965
+ },
966
+ {
967
+ "epoch": 0.33075734157650694,
968
+ "grad_norm": 2.436528444290161,
969
+ "learning_rate": 1.6695517774343124e-05,
970
+ "loss": 3.4595,
971
+ "mean_token_accuracy": 0.3885770753026009,
972
+ "num_tokens": 3378707.0,
973
+ "step": 1070
974
+ },
975
+ {
976
+ "epoch": 0.33384853168469864,
977
+ "grad_norm": 2.0806350708007812,
978
+ "learning_rate": 1.6664605873261207e-05,
979
+ "loss": 3.4576,
980
+ "mean_token_accuracy": 0.38848345205187795,
981
+ "num_tokens": 3410231.0,
982
+ "step": 1080
983
+ },
984
+ {
985
+ "epoch": 0.3369397217928903,
986
+ "grad_norm": 2.142319679260254,
987
+ "learning_rate": 1.663369397217929e-05,
988
+ "loss": 3.4644,
989
+ "mean_token_accuracy": 0.3861194223165512,
990
+ "num_tokens": 3444065.0,
991
+ "step": 1090
992
+ },
993
+ {
994
+ "epoch": 0.3400309119010819,
995
+ "grad_norm": 2.380552053451538,
996
+ "learning_rate": 1.6602782071097374e-05,
997
+ "loss": 3.3916,
998
+ "mean_token_accuracy": 0.39232398346066477,
999
+ "num_tokens": 3474974.0,
1000
+ "step": 1100
1001
+ },
1002
+ {
1003
+ "epoch": 0.34312210200927357,
1004
+ "grad_norm": 1.6462984085083008,
1005
+ "learning_rate": 1.6571870170015458e-05,
1006
+ "loss": 3.3687,
1007
+ "mean_token_accuracy": 0.3944365203380585,
1008
+ "num_tokens": 3507101.0,
1009
+ "step": 1110
1010
+ },
1011
+ {
1012
+ "epoch": 0.3462132921174652,
1013
+ "grad_norm": 4.342376232147217,
1014
+ "learning_rate": 1.654095826893354e-05,
1015
+ "loss": 3.4589,
1016
+ "mean_token_accuracy": 0.3856883034110069,
1017
+ "num_tokens": 3538893.0,
1018
+ "step": 1120
1019
+ },
1020
+ {
1021
+ "epoch": 0.34930448222565685,
1022
+ "grad_norm": 1.8314056396484375,
1023
+ "learning_rate": 1.6510046367851625e-05,
1024
+ "loss": 3.6277,
1025
+ "mean_token_accuracy": 0.3779118649661541,
1026
+ "num_tokens": 3567776.0,
1027
+ "step": 1130
1028
+ },
1029
+ {
1030
+ "epoch": 0.35239567233384855,
1031
+ "grad_norm": 1.7731289863586426,
1032
+ "learning_rate": 1.647913446676971e-05,
1033
+ "loss": 3.4375,
1034
+ "mean_token_accuracy": 0.38362068235874175,
1035
+ "num_tokens": 3600096.0,
1036
+ "step": 1140
1037
+ },
1038
+ {
1039
+ "epoch": 0.3554868624420402,
1040
+ "grad_norm": 1.4724918603897095,
1041
+ "learning_rate": 1.644822256568779e-05,
1042
+ "loss": 3.4787,
1043
+ "mean_token_accuracy": 0.39140490964055064,
1044
+ "num_tokens": 3629645.0,
1045
+ "step": 1150
1046
+ },
1047
+ {
1048
+ "epoch": 0.35857805255023184,
1049
+ "grad_norm": 2.0459768772125244,
1050
+ "learning_rate": 1.6417310664605876e-05,
1051
+ "loss": 3.4346,
1052
+ "mean_token_accuracy": 0.38677491843700407,
1053
+ "num_tokens": 3663559.0,
1054
+ "step": 1160
1055
+ },
1056
+ {
1057
+ "epoch": 0.3616692426584235,
1058
+ "grad_norm": 3.235039710998535,
1059
+ "learning_rate": 1.638639876352396e-05,
1060
+ "loss": 3.4915,
1061
+ "mean_token_accuracy": 0.38628031834959986,
1062
+ "num_tokens": 3695457.0,
1063
+ "step": 1170
1064
+ },
1065
+ {
1066
+ "epoch": 0.36476043276661513,
1067
+ "grad_norm": 1.2474194765090942,
1068
+ "learning_rate": 1.635548686244204e-05,
1069
+ "loss": 3.3999,
1070
+ "mean_token_accuracy": 0.39414023533463477,
1071
+ "num_tokens": 3725294.0,
1072
+ "step": 1180
1073
+ },
1074
+ {
1075
+ "epoch": 0.3678516228748068,
1076
+ "grad_norm": 1.3599259853363037,
1077
+ "learning_rate": 1.6324574961360123e-05,
1078
+ "loss": 3.3712,
1079
+ "mean_token_accuracy": 0.39129213988780975,
1080
+ "num_tokens": 3756722.0,
1081
+ "step": 1190
1082
+ },
1083
+ {
1084
+ "epoch": 0.37094281298299847,
1085
+ "grad_norm": 1.4785717725753784,
1086
+ "learning_rate": 1.629366306027821e-05,
1087
+ "loss": 3.3837,
1088
+ "mean_token_accuracy": 0.39773035794496536,
1089
+ "num_tokens": 3786396.0,
1090
+ "step": 1200
1091
+ },
1092
+ {
1093
+ "epoch": 0.3740340030911901,
1094
+ "grad_norm": 1.1180949211120605,
1095
+ "learning_rate": 1.6262751159196294e-05,
1096
+ "loss": 3.374,
1097
+ "mean_token_accuracy": 0.3942295677959919,
1098
+ "num_tokens": 3816940.0,
1099
+ "step": 1210
1100
+ },
1101
+ {
1102
+ "epoch": 0.37712519319938176,
1103
+ "grad_norm": 3.58443546295166,
1104
+ "learning_rate": 1.6231839258114374e-05,
1105
+ "loss": 3.3892,
1106
+ "mean_token_accuracy": 0.3918235659599304,
1107
+ "num_tokens": 3845728.0,
1108
+ "step": 1220
1109
+ },
1110
+ {
1111
+ "epoch": 0.3802163833075734,
1112
+ "grad_norm": 7.910126686096191,
1113
+ "learning_rate": 1.6200927357032458e-05,
1114
+ "loss": 3.4141,
1115
+ "mean_token_accuracy": 0.39633639603853227,
1116
+ "num_tokens": 3876251.0,
1117
+ "step": 1230
1118
+ },
1119
+ {
1120
+ "epoch": 0.38330757341576505,
1121
+ "grad_norm": 1.4737247228622437,
1122
+ "learning_rate": 1.617001545595054e-05,
1123
+ "loss": 3.3891,
1124
+ "mean_token_accuracy": 0.38747691363096237,
1125
+ "num_tokens": 3907979.0,
1126
+ "step": 1240
1127
+ },
1128
+ {
1129
+ "epoch": 0.38639876352395675,
1130
+ "grad_norm": 2.543823003768921,
1131
+ "learning_rate": 1.6139103554868625e-05,
1132
+ "loss": 3.3857,
1133
+ "mean_token_accuracy": 0.3897668160498142,
1134
+ "num_tokens": 3939556.0,
1135
+ "step": 1250
1136
+ },
1137
+ {
1138
+ "epoch": 0.3894899536321484,
1139
+ "grad_norm": 1.8837332725524902,
1140
+ "learning_rate": 1.610819165378671e-05,
1141
+ "loss": 3.3795,
1142
+ "mean_token_accuracy": 0.39769657924771307,
1143
+ "num_tokens": 3970481.0,
1144
+ "step": 1260
1145
+ },
1146
+ {
1147
+ "epoch": 0.39258114374034003,
1148
+ "grad_norm": 3.573788642883301,
1149
+ "learning_rate": 1.6077279752704792e-05,
1150
+ "loss": 3.3105,
1151
+ "mean_token_accuracy": 0.4041719429194927,
1152
+ "num_tokens": 3999666.0,
1153
+ "step": 1270
1154
+ },
1155
+ {
1156
+ "epoch": 0.3956723338485317,
1157
+ "grad_norm": 1.7824413776397705,
1158
+ "learning_rate": 1.6046367851622876e-05,
1159
+ "loss": 3.4468,
1160
+ "mean_token_accuracy": 0.3911862142384052,
1161
+ "num_tokens": 4031876.0,
1162
+ "step": 1280
1163
+ },
1164
+ {
1165
+ "epoch": 0.3987635239567233,
1166
+ "grad_norm": 1.6730329990386963,
1167
+ "learning_rate": 1.601545595054096e-05,
1168
+ "loss": 3.4193,
1169
+ "mean_token_accuracy": 0.39065720662474634,
1170
+ "num_tokens": 4064132.0,
1171
+ "step": 1290
1172
+ },
1173
+ {
1174
+ "epoch": 0.401854714064915,
1175
+ "grad_norm": 1.421411395072937,
1176
+ "learning_rate": 1.5984544049459043e-05,
1177
+ "loss": 3.4292,
1178
+ "mean_token_accuracy": 0.3871223643422127,
1179
+ "num_tokens": 4095074.0,
1180
+ "step": 1300
1181
+ },
1182
+ {
1183
+ "epoch": 0.40494590417310666,
1184
+ "grad_norm": 1.7248343229293823,
1185
+ "learning_rate": 1.5953632148377126e-05,
1186
+ "loss": 3.379,
1187
+ "mean_token_accuracy": 0.39316892698407174,
1188
+ "num_tokens": 4130858.0,
1189
+ "step": 1310
1190
+ },
1191
+ {
1192
+ "epoch": 0.4080370942812983,
1193
+ "grad_norm": 2.6556711196899414,
1194
+ "learning_rate": 1.592272024729521e-05,
1195
+ "loss": 3.4088,
1196
+ "mean_token_accuracy": 0.38396543338894845,
1197
+ "num_tokens": 4163277.0,
1198
+ "step": 1320
1199
+ },
1200
+ {
1201
+ "epoch": 0.41112828438948995,
1202
+ "grad_norm": 2.2201597690582275,
1203
+ "learning_rate": 1.5891808346213294e-05,
1204
+ "loss": 3.446,
1205
+ "mean_token_accuracy": 0.38794904500246047,
1206
+ "num_tokens": 4195806.0,
1207
+ "step": 1330
1208
+ },
1209
+ {
1210
+ "epoch": 0.4142194744976816,
1211
+ "grad_norm": 1.4001938104629517,
1212
+ "learning_rate": 1.5860896445131377e-05,
1213
+ "loss": 3.4074,
1214
+ "mean_token_accuracy": 0.3959184519946575,
1215
+ "num_tokens": 4225811.0,
1216
+ "step": 1340
1217
+ },
1218
+ {
1219
+ "epoch": 0.41731066460587324,
1220
+ "grad_norm": 1.9553899765014648,
1221
+ "learning_rate": 1.582998454404946e-05,
1222
+ "loss": 3.353,
1223
+ "mean_token_accuracy": 0.392233844101429,
1224
+ "num_tokens": 4259418.0,
1225
+ "step": 1350
1226
+ },
1227
+ {
1228
+ "epoch": 0.42040185471406494,
1229
+ "grad_norm": 2.1650876998901367,
1230
+ "learning_rate": 1.5799072642967544e-05,
1231
+ "loss": 3.3014,
1232
+ "mean_token_accuracy": 0.40291827023029325,
1233
+ "num_tokens": 4292593.0,
1234
+ "step": 1360
1235
+ },
1236
+ {
1237
+ "epoch": 0.4234930448222566,
1238
+ "grad_norm": 1.6802010536193848,
1239
+ "learning_rate": 1.5768160741885628e-05,
1240
+ "loss": 3.3615,
1241
+ "mean_token_accuracy": 0.3953541323542595,
1242
+ "num_tokens": 4325628.0,
1243
+ "step": 1370
1244
+ },
1245
+ {
1246
+ "epoch": 0.4265842349304482,
1247
+ "grad_norm": 2.8798279762268066,
1248
+ "learning_rate": 1.573724884080371e-05,
1249
+ "loss": 3.4365,
1250
+ "mean_token_accuracy": 0.39370308369398116,
1251
+ "num_tokens": 4358459.0,
1252
+ "step": 1380
1253
+ },
1254
+ {
1255
+ "epoch": 0.42967542503863987,
1256
+ "grad_norm": 1.523694634437561,
1257
+ "learning_rate": 1.5706336939721795e-05,
1258
+ "loss": 3.3153,
1259
+ "mean_token_accuracy": 0.3944700941443443,
1260
+ "num_tokens": 4390529.0,
1261
+ "step": 1390
1262
+ },
1263
+ {
1264
+ "epoch": 0.4327666151468315,
1265
+ "grad_norm": 2.0799732208251953,
1266
+ "learning_rate": 1.567542503863988e-05,
1267
+ "loss": 3.4242,
1268
+ "mean_token_accuracy": 0.389276672154665,
1269
+ "num_tokens": 4424045.0,
1270
+ "step": 1400
1271
+ },
1272
+ {
1273
+ "epoch": 0.43585780525502316,
1274
+ "grad_norm": 1.3916538953781128,
1275
+ "learning_rate": 1.5644513137557962e-05,
1276
+ "loss": 3.444,
1277
+ "mean_token_accuracy": 0.388773063570261,
1278
+ "num_tokens": 4456115.0,
1279
+ "step": 1410
1280
+ },
1281
+ {
1282
+ "epoch": 0.43894899536321486,
1283
+ "grad_norm": 11.072097778320312,
1284
+ "learning_rate": 1.5613601236476042e-05,
1285
+ "loss": 3.2926,
1286
+ "mean_token_accuracy": 0.40686351582407954,
1287
+ "num_tokens": 4490990.0,
1288
+ "step": 1420
1289
+ },
1290
+ {
1291
+ "epoch": 0.4420401854714065,
1292
+ "grad_norm": 1.8008073568344116,
1293
+ "learning_rate": 1.558268933539413e-05,
1294
+ "loss": 3.4062,
1295
+ "mean_token_accuracy": 0.3938455879688263,
1296
+ "num_tokens": 4519847.0,
1297
+ "step": 1430
1298
+ },
1299
+ {
1300
+ "epoch": 0.44513137557959814,
1301
+ "grad_norm": 4.065845012664795,
1302
+ "learning_rate": 1.5551777434312213e-05,
1303
+ "loss": 3.3776,
1304
+ "mean_token_accuracy": 0.4004744917154312,
1305
+ "num_tokens": 4551679.0,
1306
+ "step": 1440
1307
+ },
1308
+ {
1309
+ "epoch": 0.4482225656877898,
1310
+ "grad_norm": 2.4614624977111816,
1311
+ "learning_rate": 1.5520865533230297e-05,
1312
+ "loss": 3.324,
1313
+ "mean_token_accuracy": 0.4031891174614429,
1314
+ "num_tokens": 4582622.0,
1315
+ "step": 1450
1316
+ },
1317
+ {
1318
+ "epoch": 0.45131375579598143,
1319
+ "grad_norm": 2.6623904705047607,
1320
+ "learning_rate": 1.5489953632148377e-05,
1321
+ "loss": 3.414,
1322
+ "mean_token_accuracy": 0.39484291821718215,
1323
+ "num_tokens": 4613478.0,
1324
+ "step": 1460
1325
+ },
1326
+ {
1327
+ "epoch": 0.45440494590417313,
1328
+ "grad_norm": 2.342698574066162,
1329
+ "learning_rate": 1.545904173106646e-05,
1330
+ "loss": 3.4348,
1331
+ "mean_token_accuracy": 0.38681296780705454,
1332
+ "num_tokens": 4647495.0,
1333
+ "step": 1470
1334
+ },
1335
+ {
1336
+ "epoch": 0.4574961360123648,
1337
+ "grad_norm": 2.6113431453704834,
1338
+ "learning_rate": 1.5428129829984547e-05,
1339
+ "loss": 3.39,
1340
+ "mean_token_accuracy": 0.391095020622015,
1341
+ "num_tokens": 4681809.0,
1342
+ "step": 1480
1343
+ },
1344
+ {
1345
+ "epoch": 0.4605873261205564,
1346
+ "grad_norm": 1.3145941495895386,
1347
+ "learning_rate": 1.5397217928902627e-05,
1348
+ "loss": 3.3677,
1349
+ "mean_token_accuracy": 0.38954789489507674,
1350
+ "num_tokens": 4716308.0,
1351
+ "step": 1490
1352
+ },
1353
+ {
1354
+ "epoch": 0.46367851622874806,
1355
+ "grad_norm": 1.8231385946273804,
1356
+ "learning_rate": 1.536630602782071e-05,
1357
+ "loss": 3.4199,
1358
+ "mean_token_accuracy": 0.38780914843082426,
1359
+ "num_tokens": 4747835.0,
1360
+ "step": 1500
1361
+ }
1362
+ ],
1363
+ "logging_steps": 10,
1364
+ "max_steps": 6470,
1365
+ "num_input_tokens_seen": 0,
1366
+ "num_train_epochs": 2,
1367
+ "save_steps": 500,
1368
+ "stateful_callbacks": {
1369
+ "TrainerControl": {
1370
+ "args": {
1371
+ "should_epoch_stop": false,
1372
+ "should_evaluate": false,
1373
+ "should_log": false,
1374
+ "should_save": true,
1375
+ "should_training_stop": false
1376
+ },
1377
+ "attributes": {}
1378
+ }
1379
+ },
1380
+ "total_flos": 2737103442425856.0,
1381
+ "train_batch_size": 4,
1382
+ "trial_name": null,
1383
+ "trial_params": null
1384
+ }
checkpoint-1500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7b555e5b9b497f109f303fcefbaa4f3565d3b5b6b83ddc274743aea488d2cd5
3
+ size 5624
checkpoint-2000/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": 256,
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-2000/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-2000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dee4b83521c347e838db19c0b5a7c3871736cb315d5b9b052b35fc89599bedb9
3
+ size 309900448
checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb203ea8524c948f3afe53b7fdfa1dcbd09afc440dd6cd451260e7ea4473482c
3
+ size 619836730
checkpoint-2000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
checkpoint-2000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25aea70366ba38dadc180c19f3328461e10964fbb02cbb41cb1ce03c8b99c414
3
+ size 1064
checkpoint-2000/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-2000/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2000/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-2000/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-2000/trainer_state.json ADDED
@@ -0,0 +1,1834 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.6182380216383307,
6
+ "eval_steps": 500,
7
+ "global_step": 2000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0030911901081916537,
14
+ "grad_norm": 3.058769941329956,
15
+ "learning_rate": 1.9972179289026277e-05,
16
+ "loss": 5.0209,
17
+ "mean_token_accuracy": 0.21833103336393833,
18
+ "num_tokens": 28941.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.0061823802163833074,
23
+ "grad_norm": 2.485980749130249,
24
+ "learning_rate": 1.994126738794436e-05,
25
+ "loss": 4.7631,
26
+ "mean_token_accuracy": 0.23807235918939113,
27
+ "num_tokens": 60990.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.00927357032457496,
32
+ "grad_norm": 5.288544178009033,
33
+ "learning_rate": 1.9910355486862444e-05,
34
+ "loss": 4.6799,
35
+ "mean_token_accuracy": 0.24858475103974342,
36
+ "num_tokens": 90602.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.012364760432766615,
41
+ "grad_norm": 3.226719379425049,
42
+ "learning_rate": 1.9879443585780528e-05,
43
+ "loss": 4.5883,
44
+ "mean_token_accuracy": 0.2555678006261587,
45
+ "num_tokens": 121132.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.015455950540958269,
50
+ "grad_norm": 2.027597188949585,
51
+ "learning_rate": 1.984853168469861e-05,
52
+ "loss": 4.5125,
53
+ "mean_token_accuracy": 0.26637452803552153,
54
+ "num_tokens": 154244.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.01854714064914992,
59
+ "grad_norm": 1.6715962886810303,
60
+ "learning_rate": 1.9817619783616695e-05,
61
+ "loss": 4.4224,
62
+ "mean_token_accuracy": 0.27198897041380404,
63
+ "num_tokens": 185618.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.021638330757341576,
68
+ "grad_norm": 3.0472540855407715,
69
+ "learning_rate": 1.9786707882534775e-05,
70
+ "loss": 4.3464,
71
+ "mean_token_accuracy": 0.2801041007041931,
72
+ "num_tokens": 220138.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.02472952086553323,
77
+ "grad_norm": 2.74045729637146,
78
+ "learning_rate": 1.9755795981452862e-05,
79
+ "loss": 4.4569,
80
+ "mean_token_accuracy": 0.2733523309230804,
81
+ "num_tokens": 252919.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.027820710973724884,
86
+ "grad_norm": 1.9274535179138184,
87
+ "learning_rate": 1.9724884080370946e-05,
88
+ "loss": 4.2569,
89
+ "mean_token_accuracy": 0.2858310595154762,
90
+ "num_tokens": 284739.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.030911901081916538,
95
+ "grad_norm": 1.906964898109436,
96
+ "learning_rate": 1.9693972179289026e-05,
97
+ "loss": 4.2144,
98
+ "mean_token_accuracy": 0.2953581381589174,
99
+ "num_tokens": 315186.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.03400309119010819,
104
+ "grad_norm": 2.042823314666748,
105
+ "learning_rate": 1.966306027820711e-05,
106
+ "loss": 4.2595,
107
+ "mean_token_accuracy": 0.287172843888402,
108
+ "num_tokens": 347507.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.03709428129829984,
113
+ "grad_norm": 1.9889057874679565,
114
+ "learning_rate": 1.9632148377125197e-05,
115
+ "loss": 4.1643,
116
+ "mean_token_accuracy": 0.30306958928704264,
117
+ "num_tokens": 380331.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.0401854714064915,
122
+ "grad_norm": 1.7952933311462402,
123
+ "learning_rate": 1.9601236476043277e-05,
124
+ "loss": 4.1334,
125
+ "mean_token_accuracy": 0.3080880597233772,
126
+ "num_tokens": 410947.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.04327666151468315,
131
+ "grad_norm": 3.8508658409118652,
132
+ "learning_rate": 1.957032457496136e-05,
133
+ "loss": 4.0262,
134
+ "mean_token_accuracy": 0.3118089348077774,
135
+ "num_tokens": 444130.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.04636785162287481,
140
+ "grad_norm": 31.280546188354492,
141
+ "learning_rate": 1.9539412673879444e-05,
142
+ "loss": 3.9889,
143
+ "mean_token_accuracy": 0.32304659858345985,
144
+ "num_tokens": 474570.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.04945904173106646,
149
+ "grad_norm": 2.1569416522979736,
150
+ "learning_rate": 1.950850077279753e-05,
151
+ "loss": 4.1053,
152
+ "mean_token_accuracy": 0.31258094161748884,
153
+ "num_tokens": 505047.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.05255023183925812,
158
+ "grad_norm": 1.808377981185913,
159
+ "learning_rate": 1.947758887171561e-05,
160
+ "loss": 3.9711,
161
+ "mean_token_accuracy": 0.33312112018465995,
162
+ "num_tokens": 533187.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.05564142194744977,
167
+ "grad_norm": 2.8038811683654785,
168
+ "learning_rate": 1.9446676970633695e-05,
169
+ "loss": 3.961,
170
+ "mean_token_accuracy": 0.33416116759181025,
171
+ "num_tokens": 565598.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.05873261205564142,
176
+ "grad_norm": 8.930831909179688,
177
+ "learning_rate": 1.941576506955178e-05,
178
+ "loss": 4.0107,
179
+ "mean_token_accuracy": 0.3261258576065302,
180
+ "num_tokens": 598505.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.061823802163833076,
185
+ "grad_norm": 4.706038475036621,
186
+ "learning_rate": 1.9384853168469862e-05,
187
+ "loss": 3.9231,
188
+ "mean_token_accuracy": 0.33582728281617164,
189
+ "num_tokens": 630284.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.06491499227202473,
194
+ "grad_norm": 2.9916040897369385,
195
+ "learning_rate": 1.9353941267387946e-05,
196
+ "loss": 3.9407,
197
+ "mean_token_accuracy": 0.32844844460487366,
198
+ "num_tokens": 660513.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.06800618238021638,
203
+ "grad_norm": 2.763737678527832,
204
+ "learning_rate": 1.932302936630603e-05,
205
+ "loss": 4.0479,
206
+ "mean_token_accuracy": 0.3248734712600708,
207
+ "num_tokens": 693154.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.07109737248840804,
212
+ "grad_norm": 3.656487464904785,
213
+ "learning_rate": 1.9292117465224113e-05,
214
+ "loss": 3.7843,
215
+ "mean_token_accuracy": 0.34686593189835546,
216
+ "num_tokens": 724686.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.07418856259659969,
221
+ "grad_norm": 2.6140244007110596,
222
+ "learning_rate": 1.9261205564142196e-05,
223
+ "loss": 3.9261,
224
+ "mean_token_accuracy": 0.3310456670820713,
225
+ "num_tokens": 755625.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.07727975270479134,
230
+ "grad_norm": 1.595627784729004,
231
+ "learning_rate": 1.923029366306028e-05,
232
+ "loss": 3.841,
233
+ "mean_token_accuracy": 0.34514380544424056,
234
+ "num_tokens": 785478.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.080370942812983,
239
+ "grad_norm": 2.0485758781433105,
240
+ "learning_rate": 1.9199381761978363e-05,
241
+ "loss": 3.7596,
242
+ "mean_token_accuracy": 0.35417362824082377,
243
+ "num_tokens": 816351.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.08346213292117466,
248
+ "grad_norm": 1.7564281225204468,
249
+ "learning_rate": 1.9168469860896447e-05,
250
+ "loss": 3.7927,
251
+ "mean_token_accuracy": 0.3487237967550755,
252
+ "num_tokens": 849774.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.0865533230293663,
257
+ "grad_norm": 1.5662060976028442,
258
+ "learning_rate": 1.913755795981453e-05,
259
+ "loss": 3.7821,
260
+ "mean_token_accuracy": 0.3511029303073883,
261
+ "num_tokens": 881474.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.08964451313755796,
266
+ "grad_norm": 2.539433002471924,
267
+ "learning_rate": 1.9106646058732614e-05,
268
+ "loss": 3.8613,
269
+ "mean_token_accuracy": 0.3471171148121357,
270
+ "num_tokens": 913459.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.09273570324574962,
275
+ "grad_norm": 4.281046390533447,
276
+ "learning_rate": 1.9075734157650694e-05,
277
+ "loss": 3.8436,
278
+ "mean_token_accuracy": 0.3475985363125801,
279
+ "num_tokens": 947091.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.09582689335394126,
284
+ "grad_norm": 2.379791736602783,
285
+ "learning_rate": 1.904482225656878e-05,
286
+ "loss": 3.8309,
287
+ "mean_token_accuracy": 0.34195478409528735,
288
+ "num_tokens": 983471.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.09891808346213292,
293
+ "grad_norm": 2.4176697731018066,
294
+ "learning_rate": 1.9013910355486865e-05,
295
+ "loss": 3.75,
296
+ "mean_token_accuracy": 0.3633588753640652,
297
+ "num_tokens": 1011575.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.10200927357032458,
302
+ "grad_norm": 3.375523328781128,
303
+ "learning_rate": 1.898299845440495e-05,
304
+ "loss": 3.7556,
305
+ "mean_token_accuracy": 0.36073010191321375,
306
+ "num_tokens": 1042031.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.10510046367851623,
311
+ "grad_norm": 5.099122524261475,
312
+ "learning_rate": 1.895208655332303e-05,
313
+ "loss": 3.6861,
314
+ "mean_token_accuracy": 0.36301063373684883,
315
+ "num_tokens": 1074689.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.10819165378670788,
320
+ "grad_norm": 1.4775930643081665,
321
+ "learning_rate": 1.8921174652241116e-05,
322
+ "loss": 3.6744,
323
+ "mean_token_accuracy": 0.3681299857795238,
324
+ "num_tokens": 1105719.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.11128284389489954,
329
+ "grad_norm": 3.931447744369507,
330
+ "learning_rate": 1.88902627511592e-05,
331
+ "loss": 3.6248,
332
+ "mean_token_accuracy": 0.37062914595007895,
333
+ "num_tokens": 1136185.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.1143740340030912,
338
+ "grad_norm": 2.6153130531311035,
339
+ "learning_rate": 1.885935085007728e-05,
340
+ "loss": 3.6971,
341
+ "mean_token_accuracy": 0.35686987787485125,
342
+ "num_tokens": 1166461.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.11746522411128284,
347
+ "grad_norm": 3.0943849086761475,
348
+ "learning_rate": 1.8828438948995363e-05,
349
+ "loss": 3.7057,
350
+ "mean_token_accuracy": 0.36670113652944564,
351
+ "num_tokens": 1194661.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.1205564142194745,
356
+ "grad_norm": 1.760920524597168,
357
+ "learning_rate": 1.879752704791345e-05,
358
+ "loss": 3.6806,
359
+ "mean_token_accuracy": 0.366318603605032,
360
+ "num_tokens": 1224670.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.12364760432766615,
365
+ "grad_norm": 1.8976062536239624,
366
+ "learning_rate": 1.8766615146831534e-05,
367
+ "loss": 3.7318,
368
+ "mean_token_accuracy": 0.36500929966568946,
369
+ "num_tokens": 1252830.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.1267387944358578,
374
+ "grad_norm": 1.950358510017395,
375
+ "learning_rate": 1.8735703245749614e-05,
376
+ "loss": 3.7106,
377
+ "mean_token_accuracy": 0.3675060346722603,
378
+ "num_tokens": 1285546.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.12982998454404945,
383
+ "grad_norm": 2.707167148590088,
384
+ "learning_rate": 1.8704791344667697e-05,
385
+ "loss": 3.6688,
386
+ "mean_token_accuracy": 0.37156677842140196,
387
+ "num_tokens": 1316466.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.13292117465224113,
392
+ "grad_norm": 2.084510564804077,
393
+ "learning_rate": 1.8673879443585784e-05,
394
+ "loss": 3.6758,
395
+ "mean_token_accuracy": 0.3684880450367928,
396
+ "num_tokens": 1349121.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.13601236476043277,
401
+ "grad_norm": 2.2626636028289795,
402
+ "learning_rate": 1.8642967542503865e-05,
403
+ "loss": 3.7214,
404
+ "mean_token_accuracy": 0.35720510333776473,
405
+ "num_tokens": 1384366.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.1391035548686244,
410
+ "grad_norm": 2.4145290851593018,
411
+ "learning_rate": 1.8612055641421948e-05,
412
+ "loss": 3.7261,
413
+ "mean_token_accuracy": 0.36429562568664553,
414
+ "num_tokens": 1415357.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.14219474497681608,
419
+ "grad_norm": 2.7409212589263916,
420
+ "learning_rate": 1.8581143740340032e-05,
421
+ "loss": 3.7368,
422
+ "mean_token_accuracy": 0.3619597226381302,
423
+ "num_tokens": 1445641.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.14528593508500773,
428
+ "grad_norm": 4.5937275886535645,
429
+ "learning_rate": 1.8550231839258115e-05,
430
+ "loss": 3.6911,
431
+ "mean_token_accuracy": 0.3683665543794632,
432
+ "num_tokens": 1475360.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.14837712519319937,
437
+ "grad_norm": 12.00837516784668,
438
+ "learning_rate": 1.85193199381762e-05,
439
+ "loss": 3.6899,
440
+ "mean_token_accuracy": 0.3648961283266544,
441
+ "num_tokens": 1506048.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.15146831530139104,
446
+ "grad_norm": 2.5790181159973145,
447
+ "learning_rate": 1.8488408037094283e-05,
448
+ "loss": 3.6948,
449
+ "mean_token_accuracy": 0.36360194012522695,
450
+ "num_tokens": 1539998.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.1545595054095827,
455
+ "grad_norm": 1.8515182733535767,
456
+ "learning_rate": 1.8457496136012366e-05,
457
+ "loss": 3.65,
458
+ "mean_token_accuracy": 0.36655392646789553,
459
+ "num_tokens": 1573641.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.15765069551777433,
464
+ "grad_norm": 2.3960251808166504,
465
+ "learning_rate": 1.842658423493045e-05,
466
+ "loss": 3.58,
467
+ "mean_token_accuracy": 0.373273029923439,
468
+ "num_tokens": 1605367.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.160741885625966,
473
+ "grad_norm": 2.578730821609497,
474
+ "learning_rate": 1.8395672333848533e-05,
475
+ "loss": 3.6695,
476
+ "mean_token_accuracy": 0.368588350713253,
477
+ "num_tokens": 1638052.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.16383307573415765,
482
+ "grad_norm": 1.9368691444396973,
483
+ "learning_rate": 1.8364760432766617e-05,
484
+ "loss": 3.6363,
485
+ "mean_token_accuracy": 0.3681382529437542,
486
+ "num_tokens": 1672043.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.16692426584234932,
491
+ "grad_norm": 3.562593698501587,
492
+ "learning_rate": 1.83338485316847e-05,
493
+ "loss": 3.5386,
494
+ "mean_token_accuracy": 0.3801154658198357,
495
+ "num_tokens": 1704817.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.17001545595054096,
500
+ "grad_norm": 1.564929723739624,
501
+ "learning_rate": 1.8302936630602784e-05,
502
+ "loss": 3.6191,
503
+ "mean_token_accuracy": 0.36746986880898475,
504
+ "num_tokens": 1737297.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.1731066460587326,
509
+ "grad_norm": 1.4315626621246338,
510
+ "learning_rate": 1.8272024729520868e-05,
511
+ "loss": 3.592,
512
+ "mean_token_accuracy": 0.3701712526381016,
513
+ "num_tokens": 1770204.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 0.17619783616692428,
518
+ "grad_norm": 6.3183746337890625,
519
+ "learning_rate": 1.824111282843895e-05,
520
+ "loss": 3.6281,
521
+ "mean_token_accuracy": 0.37336429879069327,
522
+ "num_tokens": 1800581.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 0.17928902627511592,
527
+ "grad_norm": 1.3608078956604004,
528
+ "learning_rate": 1.8210200927357035e-05,
529
+ "loss": 3.5644,
530
+ "mean_token_accuracy": 0.37816725075244906,
531
+ "num_tokens": 1834564.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 0.18238021638330756,
536
+ "grad_norm": 11.443370819091797,
537
+ "learning_rate": 1.817928902627512e-05,
538
+ "loss": 3.6134,
539
+ "mean_token_accuracy": 0.37460487633943557,
540
+ "num_tokens": 1866113.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 0.18547140649149924,
545
+ "grad_norm": 2.1869001388549805,
546
+ "learning_rate": 1.8148377125193202e-05,
547
+ "loss": 3.6331,
548
+ "mean_token_accuracy": 0.3722210742533207,
549
+ "num_tokens": 1900627.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 0.18856259659969088,
554
+ "grad_norm": 1.8551387786865234,
555
+ "learning_rate": 1.8117465224111282e-05,
556
+ "loss": 3.5707,
557
+ "mean_token_accuracy": 0.3772866874933243,
558
+ "num_tokens": 1932988.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 0.19165378670788252,
563
+ "grad_norm": 5.528620719909668,
564
+ "learning_rate": 1.808655332302937e-05,
565
+ "loss": 3.6475,
566
+ "mean_token_accuracy": 0.3652547873556614,
567
+ "num_tokens": 1962031.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 0.1947449768160742,
572
+ "grad_norm": 3.566514253616333,
573
+ "learning_rate": 1.8055641421947453e-05,
574
+ "loss": 3.5441,
575
+ "mean_token_accuracy": 0.3836688004434109,
576
+ "num_tokens": 1995070.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 0.19783616692426584,
581
+ "grad_norm": 1.9582892656326294,
582
+ "learning_rate": 1.8024729520865533e-05,
583
+ "loss": 3.5473,
584
+ "mean_token_accuracy": 0.3757788948714733,
585
+ "num_tokens": 2024801.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 0.2009273570324575,
590
+ "grad_norm": 1.7483699321746826,
591
+ "learning_rate": 1.7993817619783616e-05,
592
+ "loss": 3.6133,
593
+ "mean_token_accuracy": 0.37615733668208123,
594
+ "num_tokens": 2054131.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 0.20401854714064915,
599
+ "grad_norm": 1.9855698347091675,
600
+ "learning_rate": 1.7962905718701703e-05,
601
+ "loss": 3.5724,
602
+ "mean_token_accuracy": 0.37552602738142016,
603
+ "num_tokens": 2086218.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 0.2071097372488408,
608
+ "grad_norm": 2.380608558654785,
609
+ "learning_rate": 1.7931993817619787e-05,
610
+ "loss": 3.5476,
611
+ "mean_token_accuracy": 0.37423405200243,
612
+ "num_tokens": 2121389.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 0.21020092735703247,
617
+ "grad_norm": 1.5300630331039429,
618
+ "learning_rate": 1.7901081916537867e-05,
619
+ "loss": 3.668,
620
+ "mean_token_accuracy": 0.3656138554215431,
621
+ "num_tokens": 2152513.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 0.2132921174652241,
626
+ "grad_norm": 1.6176111698150635,
627
+ "learning_rate": 1.787017001545595e-05,
628
+ "loss": 3.6087,
629
+ "mean_token_accuracy": 0.37140627652406694,
630
+ "num_tokens": 2184920.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 0.21638330757341576,
635
+ "grad_norm": 1.401524305343628,
636
+ "learning_rate": 1.7839258114374038e-05,
637
+ "loss": 3.4881,
638
+ "mean_token_accuracy": 0.3919842541217804,
639
+ "num_tokens": 2214174.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 0.21947449768160743,
644
+ "grad_norm": 1.4391794204711914,
645
+ "learning_rate": 1.7808346213292118e-05,
646
+ "loss": 3.5622,
647
+ "mean_token_accuracy": 0.37740132212638855,
648
+ "num_tokens": 2247593.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 0.22256568778979907,
653
+ "grad_norm": 5.364869117736816,
654
+ "learning_rate": 1.77774343122102e-05,
655
+ "loss": 3.5855,
656
+ "mean_token_accuracy": 0.37872459217905996,
657
+ "num_tokens": 2277810.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 0.22565687789799072,
662
+ "grad_norm": 1.3683693408966064,
663
+ "learning_rate": 1.7746522411128285e-05,
664
+ "loss": 3.4653,
665
+ "mean_token_accuracy": 0.38120819330215455,
666
+ "num_tokens": 2309457.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 0.2287480680061824,
671
+ "grad_norm": 1.3637374639511108,
672
+ "learning_rate": 1.771561051004637e-05,
673
+ "loss": 3.5131,
674
+ "mean_token_accuracy": 0.3864184685051441,
675
+ "num_tokens": 2339852.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 0.23183925811437403,
680
+ "grad_norm": 2.7388429641723633,
681
+ "learning_rate": 1.7684698608964452e-05,
682
+ "loss": 3.573,
683
+ "mean_token_accuracy": 0.3795412413775921,
684
+ "num_tokens": 2371444.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 0.23493044822256567,
689
+ "grad_norm": 9.798398971557617,
690
+ "learning_rate": 1.7653786707882536e-05,
691
+ "loss": 3.5441,
692
+ "mean_token_accuracy": 0.38080229982733727,
693
+ "num_tokens": 2401769.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 0.23802163833075735,
698
+ "grad_norm": 2.1416878700256348,
699
+ "learning_rate": 1.762287480680062e-05,
700
+ "loss": 3.5043,
701
+ "mean_token_accuracy": 0.3838089659810066,
702
+ "num_tokens": 2432410.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 0.241112828438949,
707
+ "grad_norm": 2.211545467376709,
708
+ "learning_rate": 1.7591962905718703e-05,
709
+ "loss": 3.5341,
710
+ "mean_token_accuracy": 0.388753118366003,
711
+ "num_tokens": 2462641.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 0.24420401854714066,
716
+ "grad_norm": 5.351387023925781,
717
+ "learning_rate": 1.7561051004636787e-05,
718
+ "loss": 3.5031,
719
+ "mean_token_accuracy": 0.3821559719741344,
720
+ "num_tokens": 2496459.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 0.2472952086553323,
725
+ "grad_norm": 2.3877508640289307,
726
+ "learning_rate": 1.753013910355487e-05,
727
+ "loss": 3.5589,
728
+ "mean_token_accuracy": 0.3789012677967548,
729
+ "num_tokens": 2526708.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 0.250386398763524,
734
+ "grad_norm": 1.2962738275527954,
735
+ "learning_rate": 1.7499227202472954e-05,
736
+ "loss": 3.505,
737
+ "mean_token_accuracy": 0.3823570780456066,
738
+ "num_tokens": 2561229.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 0.2534775888717156,
743
+ "grad_norm": 1.8147461414337158,
744
+ "learning_rate": 1.7468315301391037e-05,
745
+ "loss": 3.4861,
746
+ "mean_token_accuracy": 0.3802278622984886,
747
+ "num_tokens": 2592914.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 0.25656877897990726,
752
+ "grad_norm": 3.4690439701080322,
753
+ "learning_rate": 1.743740340030912e-05,
754
+ "loss": 3.5202,
755
+ "mean_token_accuracy": 0.38112854287028314,
756
+ "num_tokens": 2627185.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 0.2596599690880989,
761
+ "grad_norm": 4.353795051574707,
762
+ "learning_rate": 1.7406491499227205e-05,
763
+ "loss": 3.5667,
764
+ "mean_token_accuracy": 0.37200469225645066,
765
+ "num_tokens": 2660839.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 0.26275115919629055,
770
+ "grad_norm": 2.6319100856781006,
771
+ "learning_rate": 1.7375579598145288e-05,
772
+ "loss": 3.464,
773
+ "mean_token_accuracy": 0.38627258986234664,
774
+ "num_tokens": 2690042.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 0.26584234930448225,
779
+ "grad_norm": 1.6525287628173828,
780
+ "learning_rate": 1.7344667697063372e-05,
781
+ "loss": 3.5616,
782
+ "mean_token_accuracy": 0.3763453342020512,
783
+ "num_tokens": 2724354.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 0.2689335394126739,
788
+ "grad_norm": 1.9737221002578735,
789
+ "learning_rate": 1.7313755795981455e-05,
790
+ "loss": 3.5147,
791
+ "mean_token_accuracy": 0.38340551406145096,
792
+ "num_tokens": 2755591.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 0.27202472952086554,
797
+ "grad_norm": 2.5324320793151855,
798
+ "learning_rate": 1.7282843894899536e-05,
799
+ "loss": 3.4227,
800
+ "mean_token_accuracy": 0.3909419260919094,
801
+ "num_tokens": 2786566.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 0.2751159196290572,
806
+ "grad_norm": 3.3844659328460693,
807
+ "learning_rate": 1.7251931993817623e-05,
808
+ "loss": 3.5561,
809
+ "mean_token_accuracy": 0.37579271346330645,
810
+ "num_tokens": 2814171.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 0.2782071097372488,
815
+ "grad_norm": 2.7620153427124023,
816
+ "learning_rate": 1.7221020092735706e-05,
817
+ "loss": 3.5042,
818
+ "mean_token_accuracy": 0.3822973191738129,
819
+ "num_tokens": 2844402.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 0.28129829984544047,
824
+ "grad_norm": 2.089118719100952,
825
+ "learning_rate": 1.7190108191653786e-05,
826
+ "loss": 3.4583,
827
+ "mean_token_accuracy": 0.39270223304629326,
828
+ "num_tokens": 2873728.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 0.28438948995363217,
833
+ "grad_norm": 2.496480941772461,
834
+ "learning_rate": 1.715919629057187e-05,
835
+ "loss": 3.4466,
836
+ "mean_token_accuracy": 0.3897694177925587,
837
+ "num_tokens": 2905000.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 0.2874806800618238,
842
+ "grad_norm": 2.162785053253174,
843
+ "learning_rate": 1.7128284389489957e-05,
844
+ "loss": 3.4075,
845
+ "mean_token_accuracy": 0.39374888986349105,
846
+ "num_tokens": 2935706.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 0.29057187017001546,
851
+ "grad_norm": 1.5756185054779053,
852
+ "learning_rate": 1.709737248840804e-05,
853
+ "loss": 3.4396,
854
+ "mean_token_accuracy": 0.38780966177582743,
855
+ "num_tokens": 2966477.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 0.2936630602782071,
860
+ "grad_norm": 5.708625316619873,
861
+ "learning_rate": 1.706646058732612e-05,
862
+ "loss": 3.4185,
863
+ "mean_token_accuracy": 0.39326486811041833,
864
+ "num_tokens": 2996439.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 0.29675425038639874,
869
+ "grad_norm": 2.8689119815826416,
870
+ "learning_rate": 1.7035548686244204e-05,
871
+ "loss": 3.468,
872
+ "mean_token_accuracy": 0.3818407289683819,
873
+ "num_tokens": 3026332.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 0.29984544049459044,
878
+ "grad_norm": 4.020405292510986,
879
+ "learning_rate": 1.700463678516229e-05,
880
+ "loss": 3.4943,
881
+ "mean_token_accuracy": 0.3888055384159088,
882
+ "num_tokens": 3057606.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 0.3029366306027821,
887
+ "grad_norm": 1.8991873264312744,
888
+ "learning_rate": 1.697372488408037e-05,
889
+ "loss": 3.4939,
890
+ "mean_token_accuracy": 0.38414665684103966,
891
+ "num_tokens": 3089449.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 0.30602782071097373,
896
+ "grad_norm": 13.78397274017334,
897
+ "learning_rate": 1.6942812982998455e-05,
898
+ "loss": 3.5487,
899
+ "mean_token_accuracy": 0.3769320294260979,
900
+ "num_tokens": 3119537.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 0.3091190108191654,
905
+ "grad_norm": 1.599820613861084,
906
+ "learning_rate": 1.691190108191654e-05,
907
+ "loss": 3.465,
908
+ "mean_token_accuracy": 0.3884730890393257,
909
+ "num_tokens": 3151905.0,
910
+ "step": 1000
911
+ },
912
+ {
913
+ "epoch": 0.312210200927357,
914
+ "grad_norm": 6.015178680419922,
915
+ "learning_rate": 1.6880989180834622e-05,
916
+ "loss": 3.5043,
917
+ "mean_token_accuracy": 0.38035417571663854,
918
+ "num_tokens": 3184277.0,
919
+ "step": 1010
920
+ },
921
+ {
922
+ "epoch": 0.31530139103554866,
923
+ "grad_norm": 1.854093313217163,
924
+ "learning_rate": 1.6850077279752706e-05,
925
+ "loss": 3.4233,
926
+ "mean_token_accuracy": 0.389019088447094,
927
+ "num_tokens": 3217455.0,
928
+ "step": 1020
929
+ },
930
+ {
931
+ "epoch": 0.31839258114374036,
932
+ "grad_norm": 3.024531126022339,
933
+ "learning_rate": 1.681916537867079e-05,
934
+ "loss": 3.4768,
935
+ "mean_token_accuracy": 0.39151332527399063,
936
+ "num_tokens": 3251326.0,
937
+ "step": 1030
938
+ },
939
+ {
940
+ "epoch": 0.321483771251932,
941
+ "grad_norm": 1.8012628555297852,
942
+ "learning_rate": 1.6788253477588873e-05,
943
+ "loss": 3.4955,
944
+ "mean_token_accuracy": 0.3855120025575161,
945
+ "num_tokens": 3283488.0,
946
+ "step": 1040
947
+ },
948
+ {
949
+ "epoch": 0.32457496136012365,
950
+ "grad_norm": 1.4446407556533813,
951
+ "learning_rate": 1.6757341576506957e-05,
952
+ "loss": 3.4023,
953
+ "mean_token_accuracy": 0.3883494645357132,
954
+ "num_tokens": 3315970.0,
955
+ "step": 1050
956
+ },
957
+ {
958
+ "epoch": 0.3276661514683153,
959
+ "grad_norm": 2.132194995880127,
960
+ "learning_rate": 1.672642967542504e-05,
961
+ "loss": 3.5195,
962
+ "mean_token_accuracy": 0.38440208286046984,
963
+ "num_tokens": 3348645.0,
964
+ "step": 1060
965
+ },
966
+ {
967
+ "epoch": 0.33075734157650694,
968
+ "grad_norm": 2.436528444290161,
969
+ "learning_rate": 1.6695517774343124e-05,
970
+ "loss": 3.4595,
971
+ "mean_token_accuracy": 0.3885770753026009,
972
+ "num_tokens": 3378707.0,
973
+ "step": 1070
974
+ },
975
+ {
976
+ "epoch": 0.33384853168469864,
977
+ "grad_norm": 2.0806350708007812,
978
+ "learning_rate": 1.6664605873261207e-05,
979
+ "loss": 3.4576,
980
+ "mean_token_accuracy": 0.38848345205187795,
981
+ "num_tokens": 3410231.0,
982
+ "step": 1080
983
+ },
984
+ {
985
+ "epoch": 0.3369397217928903,
986
+ "grad_norm": 2.142319679260254,
987
+ "learning_rate": 1.663369397217929e-05,
988
+ "loss": 3.4644,
989
+ "mean_token_accuracy": 0.3861194223165512,
990
+ "num_tokens": 3444065.0,
991
+ "step": 1090
992
+ },
993
+ {
994
+ "epoch": 0.3400309119010819,
995
+ "grad_norm": 2.380552053451538,
996
+ "learning_rate": 1.6602782071097374e-05,
997
+ "loss": 3.3916,
998
+ "mean_token_accuracy": 0.39232398346066477,
999
+ "num_tokens": 3474974.0,
1000
+ "step": 1100
1001
+ },
1002
+ {
1003
+ "epoch": 0.34312210200927357,
1004
+ "grad_norm": 1.6462984085083008,
1005
+ "learning_rate": 1.6571870170015458e-05,
1006
+ "loss": 3.3687,
1007
+ "mean_token_accuracy": 0.3944365203380585,
1008
+ "num_tokens": 3507101.0,
1009
+ "step": 1110
1010
+ },
1011
+ {
1012
+ "epoch": 0.3462132921174652,
1013
+ "grad_norm": 4.342376232147217,
1014
+ "learning_rate": 1.654095826893354e-05,
1015
+ "loss": 3.4589,
1016
+ "mean_token_accuracy": 0.3856883034110069,
1017
+ "num_tokens": 3538893.0,
1018
+ "step": 1120
1019
+ },
1020
+ {
1021
+ "epoch": 0.34930448222565685,
1022
+ "grad_norm": 1.8314056396484375,
1023
+ "learning_rate": 1.6510046367851625e-05,
1024
+ "loss": 3.6277,
1025
+ "mean_token_accuracy": 0.3779118649661541,
1026
+ "num_tokens": 3567776.0,
1027
+ "step": 1130
1028
+ },
1029
+ {
1030
+ "epoch": 0.35239567233384855,
1031
+ "grad_norm": 1.7731289863586426,
1032
+ "learning_rate": 1.647913446676971e-05,
1033
+ "loss": 3.4375,
1034
+ "mean_token_accuracy": 0.38362068235874175,
1035
+ "num_tokens": 3600096.0,
1036
+ "step": 1140
1037
+ },
1038
+ {
1039
+ "epoch": 0.3554868624420402,
1040
+ "grad_norm": 1.4724918603897095,
1041
+ "learning_rate": 1.644822256568779e-05,
1042
+ "loss": 3.4787,
1043
+ "mean_token_accuracy": 0.39140490964055064,
1044
+ "num_tokens": 3629645.0,
1045
+ "step": 1150
1046
+ },
1047
+ {
1048
+ "epoch": 0.35857805255023184,
1049
+ "grad_norm": 2.0459768772125244,
1050
+ "learning_rate": 1.6417310664605876e-05,
1051
+ "loss": 3.4346,
1052
+ "mean_token_accuracy": 0.38677491843700407,
1053
+ "num_tokens": 3663559.0,
1054
+ "step": 1160
1055
+ },
1056
+ {
1057
+ "epoch": 0.3616692426584235,
1058
+ "grad_norm": 3.235039710998535,
1059
+ "learning_rate": 1.638639876352396e-05,
1060
+ "loss": 3.4915,
1061
+ "mean_token_accuracy": 0.38628031834959986,
1062
+ "num_tokens": 3695457.0,
1063
+ "step": 1170
1064
+ },
1065
+ {
1066
+ "epoch": 0.36476043276661513,
1067
+ "grad_norm": 1.2474194765090942,
1068
+ "learning_rate": 1.635548686244204e-05,
1069
+ "loss": 3.3999,
1070
+ "mean_token_accuracy": 0.39414023533463477,
1071
+ "num_tokens": 3725294.0,
1072
+ "step": 1180
1073
+ },
1074
+ {
1075
+ "epoch": 0.3678516228748068,
1076
+ "grad_norm": 1.3599259853363037,
1077
+ "learning_rate": 1.6324574961360123e-05,
1078
+ "loss": 3.3712,
1079
+ "mean_token_accuracy": 0.39129213988780975,
1080
+ "num_tokens": 3756722.0,
1081
+ "step": 1190
1082
+ },
1083
+ {
1084
+ "epoch": 0.37094281298299847,
1085
+ "grad_norm": 1.4785717725753784,
1086
+ "learning_rate": 1.629366306027821e-05,
1087
+ "loss": 3.3837,
1088
+ "mean_token_accuracy": 0.39773035794496536,
1089
+ "num_tokens": 3786396.0,
1090
+ "step": 1200
1091
+ },
1092
+ {
1093
+ "epoch": 0.3740340030911901,
1094
+ "grad_norm": 1.1180949211120605,
1095
+ "learning_rate": 1.6262751159196294e-05,
1096
+ "loss": 3.374,
1097
+ "mean_token_accuracy": 0.3942295677959919,
1098
+ "num_tokens": 3816940.0,
1099
+ "step": 1210
1100
+ },
1101
+ {
1102
+ "epoch": 0.37712519319938176,
1103
+ "grad_norm": 3.58443546295166,
1104
+ "learning_rate": 1.6231839258114374e-05,
1105
+ "loss": 3.3892,
1106
+ "mean_token_accuracy": 0.3918235659599304,
1107
+ "num_tokens": 3845728.0,
1108
+ "step": 1220
1109
+ },
1110
+ {
1111
+ "epoch": 0.3802163833075734,
1112
+ "grad_norm": 7.910126686096191,
1113
+ "learning_rate": 1.6200927357032458e-05,
1114
+ "loss": 3.4141,
1115
+ "mean_token_accuracy": 0.39633639603853227,
1116
+ "num_tokens": 3876251.0,
1117
+ "step": 1230
1118
+ },
1119
+ {
1120
+ "epoch": 0.38330757341576505,
1121
+ "grad_norm": 1.4737247228622437,
1122
+ "learning_rate": 1.617001545595054e-05,
1123
+ "loss": 3.3891,
1124
+ "mean_token_accuracy": 0.38747691363096237,
1125
+ "num_tokens": 3907979.0,
1126
+ "step": 1240
1127
+ },
1128
+ {
1129
+ "epoch": 0.38639876352395675,
1130
+ "grad_norm": 2.543823003768921,
1131
+ "learning_rate": 1.6139103554868625e-05,
1132
+ "loss": 3.3857,
1133
+ "mean_token_accuracy": 0.3897668160498142,
1134
+ "num_tokens": 3939556.0,
1135
+ "step": 1250
1136
+ },
1137
+ {
1138
+ "epoch": 0.3894899536321484,
1139
+ "grad_norm": 1.8837332725524902,
1140
+ "learning_rate": 1.610819165378671e-05,
1141
+ "loss": 3.3795,
1142
+ "mean_token_accuracy": 0.39769657924771307,
1143
+ "num_tokens": 3970481.0,
1144
+ "step": 1260
1145
+ },
1146
+ {
1147
+ "epoch": 0.39258114374034003,
1148
+ "grad_norm": 3.573788642883301,
1149
+ "learning_rate": 1.6077279752704792e-05,
1150
+ "loss": 3.3105,
1151
+ "mean_token_accuracy": 0.4041719429194927,
1152
+ "num_tokens": 3999666.0,
1153
+ "step": 1270
1154
+ },
1155
+ {
1156
+ "epoch": 0.3956723338485317,
1157
+ "grad_norm": 1.7824413776397705,
1158
+ "learning_rate": 1.6046367851622876e-05,
1159
+ "loss": 3.4468,
1160
+ "mean_token_accuracy": 0.3911862142384052,
1161
+ "num_tokens": 4031876.0,
1162
+ "step": 1280
1163
+ },
1164
+ {
1165
+ "epoch": 0.3987635239567233,
1166
+ "grad_norm": 1.6730329990386963,
1167
+ "learning_rate": 1.601545595054096e-05,
1168
+ "loss": 3.4193,
1169
+ "mean_token_accuracy": 0.39065720662474634,
1170
+ "num_tokens": 4064132.0,
1171
+ "step": 1290
1172
+ },
1173
+ {
1174
+ "epoch": 0.401854714064915,
1175
+ "grad_norm": 1.421411395072937,
1176
+ "learning_rate": 1.5984544049459043e-05,
1177
+ "loss": 3.4292,
1178
+ "mean_token_accuracy": 0.3871223643422127,
1179
+ "num_tokens": 4095074.0,
1180
+ "step": 1300
1181
+ },
1182
+ {
1183
+ "epoch": 0.40494590417310666,
1184
+ "grad_norm": 1.7248343229293823,
1185
+ "learning_rate": 1.5953632148377126e-05,
1186
+ "loss": 3.379,
1187
+ "mean_token_accuracy": 0.39316892698407174,
1188
+ "num_tokens": 4130858.0,
1189
+ "step": 1310
1190
+ },
1191
+ {
1192
+ "epoch": 0.4080370942812983,
1193
+ "grad_norm": 2.6556711196899414,
1194
+ "learning_rate": 1.592272024729521e-05,
1195
+ "loss": 3.4088,
1196
+ "mean_token_accuracy": 0.38396543338894845,
1197
+ "num_tokens": 4163277.0,
1198
+ "step": 1320
1199
+ },
1200
+ {
1201
+ "epoch": 0.41112828438948995,
1202
+ "grad_norm": 2.2201597690582275,
1203
+ "learning_rate": 1.5891808346213294e-05,
1204
+ "loss": 3.446,
1205
+ "mean_token_accuracy": 0.38794904500246047,
1206
+ "num_tokens": 4195806.0,
1207
+ "step": 1330
1208
+ },
1209
+ {
1210
+ "epoch": 0.4142194744976816,
1211
+ "grad_norm": 1.4001938104629517,
1212
+ "learning_rate": 1.5860896445131377e-05,
1213
+ "loss": 3.4074,
1214
+ "mean_token_accuracy": 0.3959184519946575,
1215
+ "num_tokens": 4225811.0,
1216
+ "step": 1340
1217
+ },
1218
+ {
1219
+ "epoch": 0.41731066460587324,
1220
+ "grad_norm": 1.9553899765014648,
1221
+ "learning_rate": 1.582998454404946e-05,
1222
+ "loss": 3.353,
1223
+ "mean_token_accuracy": 0.392233844101429,
1224
+ "num_tokens": 4259418.0,
1225
+ "step": 1350
1226
+ },
1227
+ {
1228
+ "epoch": 0.42040185471406494,
1229
+ "grad_norm": 2.1650876998901367,
1230
+ "learning_rate": 1.5799072642967544e-05,
1231
+ "loss": 3.3014,
1232
+ "mean_token_accuracy": 0.40291827023029325,
1233
+ "num_tokens": 4292593.0,
1234
+ "step": 1360
1235
+ },
1236
+ {
1237
+ "epoch": 0.4234930448222566,
1238
+ "grad_norm": 1.6802010536193848,
1239
+ "learning_rate": 1.5768160741885628e-05,
1240
+ "loss": 3.3615,
1241
+ "mean_token_accuracy": 0.3953541323542595,
1242
+ "num_tokens": 4325628.0,
1243
+ "step": 1370
1244
+ },
1245
+ {
1246
+ "epoch": 0.4265842349304482,
1247
+ "grad_norm": 2.8798279762268066,
1248
+ "learning_rate": 1.573724884080371e-05,
1249
+ "loss": 3.4365,
1250
+ "mean_token_accuracy": 0.39370308369398116,
1251
+ "num_tokens": 4358459.0,
1252
+ "step": 1380
1253
+ },
1254
+ {
1255
+ "epoch": 0.42967542503863987,
1256
+ "grad_norm": 1.523694634437561,
1257
+ "learning_rate": 1.5706336939721795e-05,
1258
+ "loss": 3.3153,
1259
+ "mean_token_accuracy": 0.3944700941443443,
1260
+ "num_tokens": 4390529.0,
1261
+ "step": 1390
1262
+ },
1263
+ {
1264
+ "epoch": 0.4327666151468315,
1265
+ "grad_norm": 2.0799732208251953,
1266
+ "learning_rate": 1.567542503863988e-05,
1267
+ "loss": 3.4242,
1268
+ "mean_token_accuracy": 0.389276672154665,
1269
+ "num_tokens": 4424045.0,
1270
+ "step": 1400
1271
+ },
1272
+ {
1273
+ "epoch": 0.43585780525502316,
1274
+ "grad_norm": 1.3916538953781128,
1275
+ "learning_rate": 1.5644513137557962e-05,
1276
+ "loss": 3.444,
1277
+ "mean_token_accuracy": 0.388773063570261,
1278
+ "num_tokens": 4456115.0,
1279
+ "step": 1410
1280
+ },
1281
+ {
1282
+ "epoch": 0.43894899536321486,
1283
+ "grad_norm": 11.072097778320312,
1284
+ "learning_rate": 1.5613601236476042e-05,
1285
+ "loss": 3.2926,
1286
+ "mean_token_accuracy": 0.40686351582407954,
1287
+ "num_tokens": 4490990.0,
1288
+ "step": 1420
1289
+ },
1290
+ {
1291
+ "epoch": 0.4420401854714065,
1292
+ "grad_norm": 1.8008073568344116,
1293
+ "learning_rate": 1.558268933539413e-05,
1294
+ "loss": 3.4062,
1295
+ "mean_token_accuracy": 0.3938455879688263,
1296
+ "num_tokens": 4519847.0,
1297
+ "step": 1430
1298
+ },
1299
+ {
1300
+ "epoch": 0.44513137557959814,
1301
+ "grad_norm": 4.065845012664795,
1302
+ "learning_rate": 1.5551777434312213e-05,
1303
+ "loss": 3.3776,
1304
+ "mean_token_accuracy": 0.4004744917154312,
1305
+ "num_tokens": 4551679.0,
1306
+ "step": 1440
1307
+ },
1308
+ {
1309
+ "epoch": 0.4482225656877898,
1310
+ "grad_norm": 2.4614624977111816,
1311
+ "learning_rate": 1.5520865533230297e-05,
1312
+ "loss": 3.324,
1313
+ "mean_token_accuracy": 0.4031891174614429,
1314
+ "num_tokens": 4582622.0,
1315
+ "step": 1450
1316
+ },
1317
+ {
1318
+ "epoch": 0.45131375579598143,
1319
+ "grad_norm": 2.6623904705047607,
1320
+ "learning_rate": 1.5489953632148377e-05,
1321
+ "loss": 3.414,
1322
+ "mean_token_accuracy": 0.39484291821718215,
1323
+ "num_tokens": 4613478.0,
1324
+ "step": 1460
1325
+ },
1326
+ {
1327
+ "epoch": 0.45440494590417313,
1328
+ "grad_norm": 2.342698574066162,
1329
+ "learning_rate": 1.545904173106646e-05,
1330
+ "loss": 3.4348,
1331
+ "mean_token_accuracy": 0.38681296780705454,
1332
+ "num_tokens": 4647495.0,
1333
+ "step": 1470
1334
+ },
1335
+ {
1336
+ "epoch": 0.4574961360123648,
1337
+ "grad_norm": 2.6113431453704834,
1338
+ "learning_rate": 1.5428129829984547e-05,
1339
+ "loss": 3.39,
1340
+ "mean_token_accuracy": 0.391095020622015,
1341
+ "num_tokens": 4681809.0,
1342
+ "step": 1480
1343
+ },
1344
+ {
1345
+ "epoch": 0.4605873261205564,
1346
+ "grad_norm": 1.3145941495895386,
1347
+ "learning_rate": 1.5397217928902627e-05,
1348
+ "loss": 3.3677,
1349
+ "mean_token_accuracy": 0.38954789489507674,
1350
+ "num_tokens": 4716308.0,
1351
+ "step": 1490
1352
+ },
1353
+ {
1354
+ "epoch": 0.46367851622874806,
1355
+ "grad_norm": 1.8231385946273804,
1356
+ "learning_rate": 1.536630602782071e-05,
1357
+ "loss": 3.4199,
1358
+ "mean_token_accuracy": 0.38780914843082426,
1359
+ "num_tokens": 4747835.0,
1360
+ "step": 1500
1361
+ },
1362
+ {
1363
+ "epoch": 0.4667697063369397,
1364
+ "grad_norm": 2.2594892978668213,
1365
+ "learning_rate": 1.5335394126738795e-05,
1366
+ "loss": 3.3937,
1367
+ "mean_token_accuracy": 0.3959015667438507,
1368
+ "num_tokens": 4776486.0,
1369
+ "step": 1510
1370
+ },
1371
+ {
1372
+ "epoch": 0.46986089644513135,
1373
+ "grad_norm": 3.840742349624634,
1374
+ "learning_rate": 1.5304482225656878e-05,
1375
+ "loss": 3.3087,
1376
+ "mean_token_accuracy": 0.407059845328331,
1377
+ "num_tokens": 4805943.0,
1378
+ "step": 1520
1379
+ },
1380
+ {
1381
+ "epoch": 0.47295208655332305,
1382
+ "grad_norm": 4.683995246887207,
1383
+ "learning_rate": 1.5273570324574962e-05,
1384
+ "loss": 3.3373,
1385
+ "mean_token_accuracy": 0.39871800169348715,
1386
+ "num_tokens": 4838440.0,
1387
+ "step": 1530
1388
+ },
1389
+ {
1390
+ "epoch": 0.4760432766615147,
1391
+ "grad_norm": 1.2700860500335693,
1392
+ "learning_rate": 1.5242658423493047e-05,
1393
+ "loss": 3.3339,
1394
+ "mean_token_accuracy": 0.39955019652843476,
1395
+ "num_tokens": 4870809.0,
1396
+ "step": 1540
1397
+ },
1398
+ {
1399
+ "epoch": 0.47913446676970634,
1400
+ "grad_norm": 2.15336537361145,
1401
+ "learning_rate": 1.521174652241113e-05,
1402
+ "loss": 3.3699,
1403
+ "mean_token_accuracy": 0.39904908165335656,
1404
+ "num_tokens": 4903332.0,
1405
+ "step": 1550
1406
+ },
1407
+ {
1408
+ "epoch": 0.482225656877898,
1409
+ "grad_norm": 1.299379587173462,
1410
+ "learning_rate": 1.5180834621329213e-05,
1411
+ "loss": 3.4024,
1412
+ "mean_token_accuracy": 0.3913417667150497,
1413
+ "num_tokens": 4933852.0,
1414
+ "step": 1560
1415
+ },
1416
+ {
1417
+ "epoch": 0.4853168469860896,
1418
+ "grad_norm": 1.1829091310501099,
1419
+ "learning_rate": 1.5149922720247296e-05,
1420
+ "loss": 3.3253,
1421
+ "mean_token_accuracy": 0.39646707102656364,
1422
+ "num_tokens": 4966530.0,
1423
+ "step": 1570
1424
+ },
1425
+ {
1426
+ "epoch": 0.4884080370942813,
1427
+ "grad_norm": 1.2553237676620483,
1428
+ "learning_rate": 1.511901081916538e-05,
1429
+ "loss": 3.2771,
1430
+ "mean_token_accuracy": 0.405765625834465,
1431
+ "num_tokens": 4998456.0,
1432
+ "step": 1580
1433
+ },
1434
+ {
1435
+ "epoch": 0.49149922720247297,
1436
+ "grad_norm": 2.037930488586426,
1437
+ "learning_rate": 1.5088098918083462e-05,
1438
+ "loss": 3.3754,
1439
+ "mean_token_accuracy": 0.3952351205050945,
1440
+ "num_tokens": 5028995.0,
1441
+ "step": 1590
1442
+ },
1443
+ {
1444
+ "epoch": 0.4945904173106646,
1445
+ "grad_norm": 1.090571403503418,
1446
+ "learning_rate": 1.5057187017001547e-05,
1447
+ "loss": 3.2832,
1448
+ "mean_token_accuracy": 0.4088763400912285,
1449
+ "num_tokens": 5057407.0,
1450
+ "step": 1600
1451
+ },
1452
+ {
1453
+ "epoch": 0.49768160741885625,
1454
+ "grad_norm": 2.4721672534942627,
1455
+ "learning_rate": 1.502627511591963e-05,
1456
+ "loss": 3.3535,
1457
+ "mean_token_accuracy": 0.3982778422534466,
1458
+ "num_tokens": 5088330.0,
1459
+ "step": 1610
1460
+ },
1461
+ {
1462
+ "epoch": 0.500772797527048,
1463
+ "grad_norm": 1.3971518278121948,
1464
+ "learning_rate": 1.4995363214837714e-05,
1465
+ "loss": 3.359,
1466
+ "mean_token_accuracy": 0.3978785939514637,
1467
+ "num_tokens": 5115511.0,
1468
+ "step": 1620
1469
+ },
1470
+ {
1471
+ "epoch": 0.5038639876352395,
1472
+ "grad_norm": 1.766021490097046,
1473
+ "learning_rate": 1.4964451313755796e-05,
1474
+ "loss": 3.3388,
1475
+ "mean_token_accuracy": 0.4023651979863644,
1476
+ "num_tokens": 5145560.0,
1477
+ "step": 1630
1478
+ },
1479
+ {
1480
+ "epoch": 0.5069551777434312,
1481
+ "grad_norm": 0.9848290681838989,
1482
+ "learning_rate": 1.4933539412673881e-05,
1483
+ "loss": 3.3312,
1484
+ "mean_token_accuracy": 0.4001339070498943,
1485
+ "num_tokens": 5178491.0,
1486
+ "step": 1640
1487
+ },
1488
+ {
1489
+ "epoch": 0.5100463678516228,
1490
+ "grad_norm": 5.008669376373291,
1491
+ "learning_rate": 1.4902627511591965e-05,
1492
+ "loss": 3.3395,
1493
+ "mean_token_accuracy": 0.39800570756196973,
1494
+ "num_tokens": 5209376.0,
1495
+ "step": 1650
1496
+ },
1497
+ {
1498
+ "epoch": 0.5131375579598145,
1499
+ "grad_norm": 3.3218369483947754,
1500
+ "learning_rate": 1.4871715610510047e-05,
1501
+ "loss": 3.398,
1502
+ "mean_token_accuracy": 0.3898357510566711,
1503
+ "num_tokens": 5240240.0,
1504
+ "step": 1660
1505
+ },
1506
+ {
1507
+ "epoch": 0.5162287480680062,
1508
+ "grad_norm": 1.3590713739395142,
1509
+ "learning_rate": 1.484080370942813e-05,
1510
+ "loss": 3.3246,
1511
+ "mean_token_accuracy": 0.39972665831446647,
1512
+ "num_tokens": 5269658.0,
1513
+ "step": 1670
1514
+ },
1515
+ {
1516
+ "epoch": 0.5193199381761978,
1517
+ "grad_norm": 1.647360920906067,
1518
+ "learning_rate": 1.4809891808346216e-05,
1519
+ "loss": 3.3101,
1520
+ "mean_token_accuracy": 0.4039941616356373,
1521
+ "num_tokens": 5301959.0,
1522
+ "step": 1680
1523
+ },
1524
+ {
1525
+ "epoch": 0.5224111282843895,
1526
+ "grad_norm": 1.3231589794158936,
1527
+ "learning_rate": 1.4778979907264298e-05,
1528
+ "loss": 3.2715,
1529
+ "mean_token_accuracy": 0.409288527816534,
1530
+ "num_tokens": 5332939.0,
1531
+ "step": 1690
1532
+ },
1533
+ {
1534
+ "epoch": 0.5255023183925811,
1535
+ "grad_norm": 1.8494716882705688,
1536
+ "learning_rate": 1.4748068006182381e-05,
1537
+ "loss": 3.3472,
1538
+ "mean_token_accuracy": 0.4054514840245247,
1539
+ "num_tokens": 5361660.0,
1540
+ "step": 1700
1541
+ },
1542
+ {
1543
+ "epoch": 0.5285935085007728,
1544
+ "grad_norm": 1.1796019077301025,
1545
+ "learning_rate": 1.4717156105100465e-05,
1546
+ "loss": 3.3374,
1547
+ "mean_token_accuracy": 0.40053225085139277,
1548
+ "num_tokens": 5390818.0,
1549
+ "step": 1710
1550
+ },
1551
+ {
1552
+ "epoch": 0.5316846986089645,
1553
+ "grad_norm": 1.4589906930923462,
1554
+ "learning_rate": 1.468624420401855e-05,
1555
+ "loss": 3.3532,
1556
+ "mean_token_accuracy": 0.3908236466348171,
1557
+ "num_tokens": 5423948.0,
1558
+ "step": 1720
1559
+ },
1560
+ {
1561
+ "epoch": 0.5347758887171561,
1562
+ "grad_norm": 2.48760986328125,
1563
+ "learning_rate": 1.4655332302936632e-05,
1564
+ "loss": 3.3916,
1565
+ "mean_token_accuracy": 0.39238951057195665,
1566
+ "num_tokens": 5456721.0,
1567
+ "step": 1730
1568
+ },
1569
+ {
1570
+ "epoch": 0.5378670788253478,
1571
+ "grad_norm": 1.3151116371154785,
1572
+ "learning_rate": 1.4624420401854715e-05,
1573
+ "loss": 3.3755,
1574
+ "mean_token_accuracy": 0.39836090207099917,
1575
+ "num_tokens": 5485487.0,
1576
+ "step": 1740
1577
+ },
1578
+ {
1579
+ "epoch": 0.5409582689335394,
1580
+ "grad_norm": 1.3862218856811523,
1581
+ "learning_rate": 1.4593508500772799e-05,
1582
+ "loss": 3.3972,
1583
+ "mean_token_accuracy": 0.39552291929721833,
1584
+ "num_tokens": 5517812.0,
1585
+ "step": 1750
1586
+ },
1587
+ {
1588
+ "epoch": 0.5440494590417311,
1589
+ "grad_norm": 1.4198209047317505,
1590
+ "learning_rate": 1.4562596599690881e-05,
1591
+ "loss": 3.323,
1592
+ "mean_token_accuracy": 0.4038604758679867,
1593
+ "num_tokens": 5549085.0,
1594
+ "step": 1760
1595
+ },
1596
+ {
1597
+ "epoch": 0.5471406491499228,
1598
+ "grad_norm": 2.063263416290283,
1599
+ "learning_rate": 1.4531684698608966e-05,
1600
+ "loss": 3.2286,
1601
+ "mean_token_accuracy": 0.40857273861765864,
1602
+ "num_tokens": 5580826.0,
1603
+ "step": 1770
1604
+ },
1605
+ {
1606
+ "epoch": 0.5502318392581144,
1607
+ "grad_norm": 1.5246375799179077,
1608
+ "learning_rate": 1.450077279752705e-05,
1609
+ "loss": 3.3077,
1610
+ "mean_token_accuracy": 0.3984913781285286,
1611
+ "num_tokens": 5611654.0,
1612
+ "step": 1780
1613
+ },
1614
+ {
1615
+ "epoch": 0.5533230293663061,
1616
+ "grad_norm": 3.1407299041748047,
1617
+ "learning_rate": 1.4469860896445132e-05,
1618
+ "loss": 3.3631,
1619
+ "mean_token_accuracy": 0.4005904957652092,
1620
+ "num_tokens": 5643972.0,
1621
+ "step": 1790
1622
+ },
1623
+ {
1624
+ "epoch": 0.5564142194744977,
1625
+ "grad_norm": 1.1959761381149292,
1626
+ "learning_rate": 1.4438948995363215e-05,
1627
+ "loss": 3.3863,
1628
+ "mean_token_accuracy": 0.39492699652910235,
1629
+ "num_tokens": 5674824.0,
1630
+ "step": 1800
1631
+ },
1632
+ {
1633
+ "epoch": 0.5595054095826894,
1634
+ "grad_norm": 1.6108070611953735,
1635
+ "learning_rate": 1.44080370942813e-05,
1636
+ "loss": 3.3524,
1637
+ "mean_token_accuracy": 0.3943831264972687,
1638
+ "num_tokens": 5710374.0,
1639
+ "step": 1810
1640
+ },
1641
+ {
1642
+ "epoch": 0.5625965996908809,
1643
+ "grad_norm": 2.3074188232421875,
1644
+ "learning_rate": 1.4377125193199384e-05,
1645
+ "loss": 3.4149,
1646
+ "mean_token_accuracy": 0.3902184680104256,
1647
+ "num_tokens": 5740450.0,
1648
+ "step": 1820
1649
+ },
1650
+ {
1651
+ "epoch": 0.5656877897990726,
1652
+ "grad_norm": 1.2210988998413086,
1653
+ "learning_rate": 1.4346213292117466e-05,
1654
+ "loss": 3.368,
1655
+ "mean_token_accuracy": 0.3951147675514221,
1656
+ "num_tokens": 5769865.0,
1657
+ "step": 1830
1658
+ },
1659
+ {
1660
+ "epoch": 0.5687789799072643,
1661
+ "grad_norm": 1.301492691040039,
1662
+ "learning_rate": 1.431530139103555e-05,
1663
+ "loss": 3.4011,
1664
+ "mean_token_accuracy": 0.39827719777822496,
1665
+ "num_tokens": 5798647.0,
1666
+ "step": 1840
1667
+ },
1668
+ {
1669
+ "epoch": 0.5718701700154559,
1670
+ "grad_norm": 1.8520913124084473,
1671
+ "learning_rate": 1.4284389489953633e-05,
1672
+ "loss": 3.3297,
1673
+ "mean_token_accuracy": 0.401262603700161,
1674
+ "num_tokens": 5830176.0,
1675
+ "step": 1850
1676
+ },
1677
+ {
1678
+ "epoch": 0.5749613601236476,
1679
+ "grad_norm": 1.7829524278640747,
1680
+ "learning_rate": 1.4253477588871715e-05,
1681
+ "loss": 3.2177,
1682
+ "mean_token_accuracy": 0.4113538973033428,
1683
+ "num_tokens": 5860769.0,
1684
+ "step": 1860
1685
+ },
1686
+ {
1687
+ "epoch": 0.5780525502318392,
1688
+ "grad_norm": 1.5205345153808594,
1689
+ "learning_rate": 1.42225656877898e-05,
1690
+ "loss": 3.4594,
1691
+ "mean_token_accuracy": 0.3904502220451832,
1692
+ "num_tokens": 5896213.0,
1693
+ "step": 1870
1694
+ },
1695
+ {
1696
+ "epoch": 0.5811437403400309,
1697
+ "grad_norm": 1.492475986480713,
1698
+ "learning_rate": 1.4191653786707884e-05,
1699
+ "loss": 3.2639,
1700
+ "mean_token_accuracy": 0.40337538048624993,
1701
+ "num_tokens": 5928539.0,
1702
+ "step": 1880
1703
+ },
1704
+ {
1705
+ "epoch": 0.5842349304482226,
1706
+ "grad_norm": 2.07590651512146,
1707
+ "learning_rate": 1.4160741885625968e-05,
1708
+ "loss": 3.3918,
1709
+ "mean_token_accuracy": 0.39563274309039115,
1710
+ "num_tokens": 5960442.0,
1711
+ "step": 1890
1712
+ },
1713
+ {
1714
+ "epoch": 0.5873261205564142,
1715
+ "grad_norm": 2.6959567070007324,
1716
+ "learning_rate": 1.412982998454405e-05,
1717
+ "loss": 3.3494,
1718
+ "mean_token_accuracy": 0.40076613500714303,
1719
+ "num_tokens": 5990404.0,
1720
+ "step": 1900
1721
+ },
1722
+ {
1723
+ "epoch": 0.5904173106646059,
1724
+ "grad_norm": 1.291227102279663,
1725
+ "learning_rate": 1.4098918083462135e-05,
1726
+ "loss": 3.3794,
1727
+ "mean_token_accuracy": 0.3960807867348194,
1728
+ "num_tokens": 6023418.0,
1729
+ "step": 1910
1730
+ },
1731
+ {
1732
+ "epoch": 0.5935085007727975,
1733
+ "grad_norm": 3.709761381149292,
1734
+ "learning_rate": 1.4068006182380218e-05,
1735
+ "loss": 3.3441,
1736
+ "mean_token_accuracy": 0.3981980659067631,
1737
+ "num_tokens": 6051604.0,
1738
+ "step": 1920
1739
+ },
1740
+ {
1741
+ "epoch": 0.5965996908809892,
1742
+ "grad_norm": 1.5294419527053833,
1743
+ "learning_rate": 1.40370942812983e-05,
1744
+ "loss": 3.3821,
1745
+ "mean_token_accuracy": 0.40094061717391016,
1746
+ "num_tokens": 6081609.0,
1747
+ "step": 1930
1748
+ },
1749
+ {
1750
+ "epoch": 0.5996908809891809,
1751
+ "grad_norm": 1.792324185371399,
1752
+ "learning_rate": 1.4006182380216384e-05,
1753
+ "loss": 3.3228,
1754
+ "mean_token_accuracy": 0.40257288739085195,
1755
+ "num_tokens": 6113487.0,
1756
+ "step": 1940
1757
+ },
1758
+ {
1759
+ "epoch": 0.6027820710973725,
1760
+ "grad_norm": 1.3564989566802979,
1761
+ "learning_rate": 1.3975270479134469e-05,
1762
+ "loss": 3.3603,
1763
+ "mean_token_accuracy": 0.3971732698380947,
1764
+ "num_tokens": 6143546.0,
1765
+ "step": 1950
1766
+ },
1767
+ {
1768
+ "epoch": 0.6058732612055642,
1769
+ "grad_norm": 1.8166826963424683,
1770
+ "learning_rate": 1.3944358578052551e-05,
1771
+ "loss": 3.3511,
1772
+ "mean_token_accuracy": 0.3947428591549397,
1773
+ "num_tokens": 6177302.0,
1774
+ "step": 1960
1775
+ },
1776
+ {
1777
+ "epoch": 0.6089644513137558,
1778
+ "grad_norm": 2.3519530296325684,
1779
+ "learning_rate": 1.3913446676970635e-05,
1780
+ "loss": 3.3258,
1781
+ "mean_token_accuracy": 0.40014824345707894,
1782
+ "num_tokens": 6209913.0,
1783
+ "step": 1970
1784
+ },
1785
+ {
1786
+ "epoch": 0.6120556414219475,
1787
+ "grad_norm": 1.9432330131530762,
1788
+ "learning_rate": 1.3882534775888718e-05,
1789
+ "loss": 3.2683,
1790
+ "mean_token_accuracy": 0.40405927672982217,
1791
+ "num_tokens": 6241773.0,
1792
+ "step": 1980
1793
+ },
1794
+ {
1795
+ "epoch": 0.615146831530139,
1796
+ "grad_norm": 1.2190839052200317,
1797
+ "learning_rate": 1.3851622874806803e-05,
1798
+ "loss": 3.3476,
1799
+ "mean_token_accuracy": 0.4006108805537224,
1800
+ "num_tokens": 6274930.0,
1801
+ "step": 1990
1802
+ },
1803
+ {
1804
+ "epoch": 0.6182380216383307,
1805
+ "grad_norm": 1.395822525024414,
1806
+ "learning_rate": 1.3820710973724885e-05,
1807
+ "loss": 3.2706,
1808
+ "mean_token_accuracy": 0.40902155488729475,
1809
+ "num_tokens": 6306366.0,
1810
+ "step": 2000
1811
+ }
1812
+ ],
1813
+ "logging_steps": 10,
1814
+ "max_steps": 6470,
1815
+ "num_input_tokens_seen": 0,
1816
+ "num_train_epochs": 2,
1817
+ "save_steps": 500,
1818
+ "stateful_callbacks": {
1819
+ "TrainerControl": {
1820
+ "args": {
1821
+ "should_epoch_stop": false,
1822
+ "should_evaluate": false,
1823
+ "should_log": false,
1824
+ "should_save": true,
1825
+ "should_training_stop": false
1826
+ },
1827
+ "attributes": {}
1828
+ }
1829
+ },
1830
+ "total_flos": 3634287971014656.0,
1831
+ "train_batch_size": 4,
1832
+ "trial_name": null,
1833
+ "trial_params": null
1834
+ }
checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7b555e5b9b497f109f303fcefbaa4f3565d3b5b6b83ddc274743aea488d2cd5
3
+ size 5624
checkpoint-2500/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": 256,
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-2500/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-2500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1970342e3c31f012b18291ff98cbf516272f543957402fc43c48414cc8024177
3
+ size 309900448
checkpoint-2500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f49ca0491df1b619b73561f32487fc2de407f44f104d5c65a777d00ba6915d0
3
+ size 619836730
checkpoint-2500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
checkpoint-2500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41d5c27cef2c430cccafa9221ce8e1bd4692c02232645979479d2884b78b69ed
3
+ size 1064
checkpoint-2500/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-2500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2500/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-2500/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-2500/trainer_state.json ADDED
@@ -0,0 +1,2284 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.7727975270479135,
6
+ "eval_steps": 500,
7
+ "global_step": 2500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0030911901081916537,
14
+ "grad_norm": 3.058769941329956,
15
+ "learning_rate": 1.9972179289026277e-05,
16
+ "loss": 5.0209,
17
+ "mean_token_accuracy": 0.21833103336393833,
18
+ "num_tokens": 28941.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.0061823802163833074,
23
+ "grad_norm": 2.485980749130249,
24
+ "learning_rate": 1.994126738794436e-05,
25
+ "loss": 4.7631,
26
+ "mean_token_accuracy": 0.23807235918939113,
27
+ "num_tokens": 60990.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.00927357032457496,
32
+ "grad_norm": 5.288544178009033,
33
+ "learning_rate": 1.9910355486862444e-05,
34
+ "loss": 4.6799,
35
+ "mean_token_accuracy": 0.24858475103974342,
36
+ "num_tokens": 90602.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.012364760432766615,
41
+ "grad_norm": 3.226719379425049,
42
+ "learning_rate": 1.9879443585780528e-05,
43
+ "loss": 4.5883,
44
+ "mean_token_accuracy": 0.2555678006261587,
45
+ "num_tokens": 121132.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.015455950540958269,
50
+ "grad_norm": 2.027597188949585,
51
+ "learning_rate": 1.984853168469861e-05,
52
+ "loss": 4.5125,
53
+ "mean_token_accuracy": 0.26637452803552153,
54
+ "num_tokens": 154244.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.01854714064914992,
59
+ "grad_norm": 1.6715962886810303,
60
+ "learning_rate": 1.9817619783616695e-05,
61
+ "loss": 4.4224,
62
+ "mean_token_accuracy": 0.27198897041380404,
63
+ "num_tokens": 185618.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.021638330757341576,
68
+ "grad_norm": 3.0472540855407715,
69
+ "learning_rate": 1.9786707882534775e-05,
70
+ "loss": 4.3464,
71
+ "mean_token_accuracy": 0.2801041007041931,
72
+ "num_tokens": 220138.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.02472952086553323,
77
+ "grad_norm": 2.74045729637146,
78
+ "learning_rate": 1.9755795981452862e-05,
79
+ "loss": 4.4569,
80
+ "mean_token_accuracy": 0.2733523309230804,
81
+ "num_tokens": 252919.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.027820710973724884,
86
+ "grad_norm": 1.9274535179138184,
87
+ "learning_rate": 1.9724884080370946e-05,
88
+ "loss": 4.2569,
89
+ "mean_token_accuracy": 0.2858310595154762,
90
+ "num_tokens": 284739.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.030911901081916538,
95
+ "grad_norm": 1.906964898109436,
96
+ "learning_rate": 1.9693972179289026e-05,
97
+ "loss": 4.2144,
98
+ "mean_token_accuracy": 0.2953581381589174,
99
+ "num_tokens": 315186.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.03400309119010819,
104
+ "grad_norm": 2.042823314666748,
105
+ "learning_rate": 1.966306027820711e-05,
106
+ "loss": 4.2595,
107
+ "mean_token_accuracy": 0.287172843888402,
108
+ "num_tokens": 347507.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.03709428129829984,
113
+ "grad_norm": 1.9889057874679565,
114
+ "learning_rate": 1.9632148377125197e-05,
115
+ "loss": 4.1643,
116
+ "mean_token_accuracy": 0.30306958928704264,
117
+ "num_tokens": 380331.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.0401854714064915,
122
+ "grad_norm": 1.7952933311462402,
123
+ "learning_rate": 1.9601236476043277e-05,
124
+ "loss": 4.1334,
125
+ "mean_token_accuracy": 0.3080880597233772,
126
+ "num_tokens": 410947.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.04327666151468315,
131
+ "grad_norm": 3.8508658409118652,
132
+ "learning_rate": 1.957032457496136e-05,
133
+ "loss": 4.0262,
134
+ "mean_token_accuracy": 0.3118089348077774,
135
+ "num_tokens": 444130.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.04636785162287481,
140
+ "grad_norm": 31.280546188354492,
141
+ "learning_rate": 1.9539412673879444e-05,
142
+ "loss": 3.9889,
143
+ "mean_token_accuracy": 0.32304659858345985,
144
+ "num_tokens": 474570.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.04945904173106646,
149
+ "grad_norm": 2.1569416522979736,
150
+ "learning_rate": 1.950850077279753e-05,
151
+ "loss": 4.1053,
152
+ "mean_token_accuracy": 0.31258094161748884,
153
+ "num_tokens": 505047.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.05255023183925812,
158
+ "grad_norm": 1.808377981185913,
159
+ "learning_rate": 1.947758887171561e-05,
160
+ "loss": 3.9711,
161
+ "mean_token_accuracy": 0.33312112018465995,
162
+ "num_tokens": 533187.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.05564142194744977,
167
+ "grad_norm": 2.8038811683654785,
168
+ "learning_rate": 1.9446676970633695e-05,
169
+ "loss": 3.961,
170
+ "mean_token_accuracy": 0.33416116759181025,
171
+ "num_tokens": 565598.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.05873261205564142,
176
+ "grad_norm": 8.930831909179688,
177
+ "learning_rate": 1.941576506955178e-05,
178
+ "loss": 4.0107,
179
+ "mean_token_accuracy": 0.3261258576065302,
180
+ "num_tokens": 598505.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.061823802163833076,
185
+ "grad_norm": 4.706038475036621,
186
+ "learning_rate": 1.9384853168469862e-05,
187
+ "loss": 3.9231,
188
+ "mean_token_accuracy": 0.33582728281617164,
189
+ "num_tokens": 630284.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.06491499227202473,
194
+ "grad_norm": 2.9916040897369385,
195
+ "learning_rate": 1.9353941267387946e-05,
196
+ "loss": 3.9407,
197
+ "mean_token_accuracy": 0.32844844460487366,
198
+ "num_tokens": 660513.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.06800618238021638,
203
+ "grad_norm": 2.763737678527832,
204
+ "learning_rate": 1.932302936630603e-05,
205
+ "loss": 4.0479,
206
+ "mean_token_accuracy": 0.3248734712600708,
207
+ "num_tokens": 693154.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.07109737248840804,
212
+ "grad_norm": 3.656487464904785,
213
+ "learning_rate": 1.9292117465224113e-05,
214
+ "loss": 3.7843,
215
+ "mean_token_accuracy": 0.34686593189835546,
216
+ "num_tokens": 724686.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.07418856259659969,
221
+ "grad_norm": 2.6140244007110596,
222
+ "learning_rate": 1.9261205564142196e-05,
223
+ "loss": 3.9261,
224
+ "mean_token_accuracy": 0.3310456670820713,
225
+ "num_tokens": 755625.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.07727975270479134,
230
+ "grad_norm": 1.595627784729004,
231
+ "learning_rate": 1.923029366306028e-05,
232
+ "loss": 3.841,
233
+ "mean_token_accuracy": 0.34514380544424056,
234
+ "num_tokens": 785478.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.080370942812983,
239
+ "grad_norm": 2.0485758781433105,
240
+ "learning_rate": 1.9199381761978363e-05,
241
+ "loss": 3.7596,
242
+ "mean_token_accuracy": 0.35417362824082377,
243
+ "num_tokens": 816351.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.08346213292117466,
248
+ "grad_norm": 1.7564281225204468,
249
+ "learning_rate": 1.9168469860896447e-05,
250
+ "loss": 3.7927,
251
+ "mean_token_accuracy": 0.3487237967550755,
252
+ "num_tokens": 849774.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.0865533230293663,
257
+ "grad_norm": 1.5662060976028442,
258
+ "learning_rate": 1.913755795981453e-05,
259
+ "loss": 3.7821,
260
+ "mean_token_accuracy": 0.3511029303073883,
261
+ "num_tokens": 881474.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.08964451313755796,
266
+ "grad_norm": 2.539433002471924,
267
+ "learning_rate": 1.9106646058732614e-05,
268
+ "loss": 3.8613,
269
+ "mean_token_accuracy": 0.3471171148121357,
270
+ "num_tokens": 913459.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.09273570324574962,
275
+ "grad_norm": 4.281046390533447,
276
+ "learning_rate": 1.9075734157650694e-05,
277
+ "loss": 3.8436,
278
+ "mean_token_accuracy": 0.3475985363125801,
279
+ "num_tokens": 947091.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.09582689335394126,
284
+ "grad_norm": 2.379791736602783,
285
+ "learning_rate": 1.904482225656878e-05,
286
+ "loss": 3.8309,
287
+ "mean_token_accuracy": 0.34195478409528735,
288
+ "num_tokens": 983471.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.09891808346213292,
293
+ "grad_norm": 2.4176697731018066,
294
+ "learning_rate": 1.9013910355486865e-05,
295
+ "loss": 3.75,
296
+ "mean_token_accuracy": 0.3633588753640652,
297
+ "num_tokens": 1011575.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.10200927357032458,
302
+ "grad_norm": 3.375523328781128,
303
+ "learning_rate": 1.898299845440495e-05,
304
+ "loss": 3.7556,
305
+ "mean_token_accuracy": 0.36073010191321375,
306
+ "num_tokens": 1042031.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.10510046367851623,
311
+ "grad_norm": 5.099122524261475,
312
+ "learning_rate": 1.895208655332303e-05,
313
+ "loss": 3.6861,
314
+ "mean_token_accuracy": 0.36301063373684883,
315
+ "num_tokens": 1074689.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.10819165378670788,
320
+ "grad_norm": 1.4775930643081665,
321
+ "learning_rate": 1.8921174652241116e-05,
322
+ "loss": 3.6744,
323
+ "mean_token_accuracy": 0.3681299857795238,
324
+ "num_tokens": 1105719.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.11128284389489954,
329
+ "grad_norm": 3.931447744369507,
330
+ "learning_rate": 1.88902627511592e-05,
331
+ "loss": 3.6248,
332
+ "mean_token_accuracy": 0.37062914595007895,
333
+ "num_tokens": 1136185.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.1143740340030912,
338
+ "grad_norm": 2.6153130531311035,
339
+ "learning_rate": 1.885935085007728e-05,
340
+ "loss": 3.6971,
341
+ "mean_token_accuracy": 0.35686987787485125,
342
+ "num_tokens": 1166461.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.11746522411128284,
347
+ "grad_norm": 3.0943849086761475,
348
+ "learning_rate": 1.8828438948995363e-05,
349
+ "loss": 3.7057,
350
+ "mean_token_accuracy": 0.36670113652944564,
351
+ "num_tokens": 1194661.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.1205564142194745,
356
+ "grad_norm": 1.760920524597168,
357
+ "learning_rate": 1.879752704791345e-05,
358
+ "loss": 3.6806,
359
+ "mean_token_accuracy": 0.366318603605032,
360
+ "num_tokens": 1224670.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.12364760432766615,
365
+ "grad_norm": 1.8976062536239624,
366
+ "learning_rate": 1.8766615146831534e-05,
367
+ "loss": 3.7318,
368
+ "mean_token_accuracy": 0.36500929966568946,
369
+ "num_tokens": 1252830.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.1267387944358578,
374
+ "grad_norm": 1.950358510017395,
375
+ "learning_rate": 1.8735703245749614e-05,
376
+ "loss": 3.7106,
377
+ "mean_token_accuracy": 0.3675060346722603,
378
+ "num_tokens": 1285546.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.12982998454404945,
383
+ "grad_norm": 2.707167148590088,
384
+ "learning_rate": 1.8704791344667697e-05,
385
+ "loss": 3.6688,
386
+ "mean_token_accuracy": 0.37156677842140196,
387
+ "num_tokens": 1316466.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.13292117465224113,
392
+ "grad_norm": 2.084510564804077,
393
+ "learning_rate": 1.8673879443585784e-05,
394
+ "loss": 3.6758,
395
+ "mean_token_accuracy": 0.3684880450367928,
396
+ "num_tokens": 1349121.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.13601236476043277,
401
+ "grad_norm": 2.2626636028289795,
402
+ "learning_rate": 1.8642967542503865e-05,
403
+ "loss": 3.7214,
404
+ "mean_token_accuracy": 0.35720510333776473,
405
+ "num_tokens": 1384366.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.1391035548686244,
410
+ "grad_norm": 2.4145290851593018,
411
+ "learning_rate": 1.8612055641421948e-05,
412
+ "loss": 3.7261,
413
+ "mean_token_accuracy": 0.36429562568664553,
414
+ "num_tokens": 1415357.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.14219474497681608,
419
+ "grad_norm": 2.7409212589263916,
420
+ "learning_rate": 1.8581143740340032e-05,
421
+ "loss": 3.7368,
422
+ "mean_token_accuracy": 0.3619597226381302,
423
+ "num_tokens": 1445641.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.14528593508500773,
428
+ "grad_norm": 4.5937275886535645,
429
+ "learning_rate": 1.8550231839258115e-05,
430
+ "loss": 3.6911,
431
+ "mean_token_accuracy": 0.3683665543794632,
432
+ "num_tokens": 1475360.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.14837712519319937,
437
+ "grad_norm": 12.00837516784668,
438
+ "learning_rate": 1.85193199381762e-05,
439
+ "loss": 3.6899,
440
+ "mean_token_accuracy": 0.3648961283266544,
441
+ "num_tokens": 1506048.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.15146831530139104,
446
+ "grad_norm": 2.5790181159973145,
447
+ "learning_rate": 1.8488408037094283e-05,
448
+ "loss": 3.6948,
449
+ "mean_token_accuracy": 0.36360194012522695,
450
+ "num_tokens": 1539998.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.1545595054095827,
455
+ "grad_norm": 1.8515182733535767,
456
+ "learning_rate": 1.8457496136012366e-05,
457
+ "loss": 3.65,
458
+ "mean_token_accuracy": 0.36655392646789553,
459
+ "num_tokens": 1573641.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.15765069551777433,
464
+ "grad_norm": 2.3960251808166504,
465
+ "learning_rate": 1.842658423493045e-05,
466
+ "loss": 3.58,
467
+ "mean_token_accuracy": 0.373273029923439,
468
+ "num_tokens": 1605367.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.160741885625966,
473
+ "grad_norm": 2.578730821609497,
474
+ "learning_rate": 1.8395672333848533e-05,
475
+ "loss": 3.6695,
476
+ "mean_token_accuracy": 0.368588350713253,
477
+ "num_tokens": 1638052.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.16383307573415765,
482
+ "grad_norm": 1.9368691444396973,
483
+ "learning_rate": 1.8364760432766617e-05,
484
+ "loss": 3.6363,
485
+ "mean_token_accuracy": 0.3681382529437542,
486
+ "num_tokens": 1672043.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.16692426584234932,
491
+ "grad_norm": 3.562593698501587,
492
+ "learning_rate": 1.83338485316847e-05,
493
+ "loss": 3.5386,
494
+ "mean_token_accuracy": 0.3801154658198357,
495
+ "num_tokens": 1704817.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.17001545595054096,
500
+ "grad_norm": 1.564929723739624,
501
+ "learning_rate": 1.8302936630602784e-05,
502
+ "loss": 3.6191,
503
+ "mean_token_accuracy": 0.36746986880898475,
504
+ "num_tokens": 1737297.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.1731066460587326,
509
+ "grad_norm": 1.4315626621246338,
510
+ "learning_rate": 1.8272024729520868e-05,
511
+ "loss": 3.592,
512
+ "mean_token_accuracy": 0.3701712526381016,
513
+ "num_tokens": 1770204.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 0.17619783616692428,
518
+ "grad_norm": 6.3183746337890625,
519
+ "learning_rate": 1.824111282843895e-05,
520
+ "loss": 3.6281,
521
+ "mean_token_accuracy": 0.37336429879069327,
522
+ "num_tokens": 1800581.0,
523
+ "step": 570
524
+ },
525
+ {
526
+ "epoch": 0.17928902627511592,
527
+ "grad_norm": 1.3608078956604004,
528
+ "learning_rate": 1.8210200927357035e-05,
529
+ "loss": 3.5644,
530
+ "mean_token_accuracy": 0.37816725075244906,
531
+ "num_tokens": 1834564.0,
532
+ "step": 580
533
+ },
534
+ {
535
+ "epoch": 0.18238021638330756,
536
+ "grad_norm": 11.443370819091797,
537
+ "learning_rate": 1.817928902627512e-05,
538
+ "loss": 3.6134,
539
+ "mean_token_accuracy": 0.37460487633943557,
540
+ "num_tokens": 1866113.0,
541
+ "step": 590
542
+ },
543
+ {
544
+ "epoch": 0.18547140649149924,
545
+ "grad_norm": 2.1869001388549805,
546
+ "learning_rate": 1.8148377125193202e-05,
547
+ "loss": 3.6331,
548
+ "mean_token_accuracy": 0.3722210742533207,
549
+ "num_tokens": 1900627.0,
550
+ "step": 600
551
+ },
552
+ {
553
+ "epoch": 0.18856259659969088,
554
+ "grad_norm": 1.8551387786865234,
555
+ "learning_rate": 1.8117465224111282e-05,
556
+ "loss": 3.5707,
557
+ "mean_token_accuracy": 0.3772866874933243,
558
+ "num_tokens": 1932988.0,
559
+ "step": 610
560
+ },
561
+ {
562
+ "epoch": 0.19165378670788252,
563
+ "grad_norm": 5.528620719909668,
564
+ "learning_rate": 1.808655332302937e-05,
565
+ "loss": 3.6475,
566
+ "mean_token_accuracy": 0.3652547873556614,
567
+ "num_tokens": 1962031.0,
568
+ "step": 620
569
+ },
570
+ {
571
+ "epoch": 0.1947449768160742,
572
+ "grad_norm": 3.566514253616333,
573
+ "learning_rate": 1.8055641421947453e-05,
574
+ "loss": 3.5441,
575
+ "mean_token_accuracy": 0.3836688004434109,
576
+ "num_tokens": 1995070.0,
577
+ "step": 630
578
+ },
579
+ {
580
+ "epoch": 0.19783616692426584,
581
+ "grad_norm": 1.9582892656326294,
582
+ "learning_rate": 1.8024729520865533e-05,
583
+ "loss": 3.5473,
584
+ "mean_token_accuracy": 0.3757788948714733,
585
+ "num_tokens": 2024801.0,
586
+ "step": 640
587
+ },
588
+ {
589
+ "epoch": 0.2009273570324575,
590
+ "grad_norm": 1.7483699321746826,
591
+ "learning_rate": 1.7993817619783616e-05,
592
+ "loss": 3.6133,
593
+ "mean_token_accuracy": 0.37615733668208123,
594
+ "num_tokens": 2054131.0,
595
+ "step": 650
596
+ },
597
+ {
598
+ "epoch": 0.20401854714064915,
599
+ "grad_norm": 1.9855698347091675,
600
+ "learning_rate": 1.7962905718701703e-05,
601
+ "loss": 3.5724,
602
+ "mean_token_accuracy": 0.37552602738142016,
603
+ "num_tokens": 2086218.0,
604
+ "step": 660
605
+ },
606
+ {
607
+ "epoch": 0.2071097372488408,
608
+ "grad_norm": 2.380608558654785,
609
+ "learning_rate": 1.7931993817619787e-05,
610
+ "loss": 3.5476,
611
+ "mean_token_accuracy": 0.37423405200243,
612
+ "num_tokens": 2121389.0,
613
+ "step": 670
614
+ },
615
+ {
616
+ "epoch": 0.21020092735703247,
617
+ "grad_norm": 1.5300630331039429,
618
+ "learning_rate": 1.7901081916537867e-05,
619
+ "loss": 3.668,
620
+ "mean_token_accuracy": 0.3656138554215431,
621
+ "num_tokens": 2152513.0,
622
+ "step": 680
623
+ },
624
+ {
625
+ "epoch": 0.2132921174652241,
626
+ "grad_norm": 1.6176111698150635,
627
+ "learning_rate": 1.787017001545595e-05,
628
+ "loss": 3.6087,
629
+ "mean_token_accuracy": 0.37140627652406694,
630
+ "num_tokens": 2184920.0,
631
+ "step": 690
632
+ },
633
+ {
634
+ "epoch": 0.21638330757341576,
635
+ "grad_norm": 1.401524305343628,
636
+ "learning_rate": 1.7839258114374038e-05,
637
+ "loss": 3.4881,
638
+ "mean_token_accuracy": 0.3919842541217804,
639
+ "num_tokens": 2214174.0,
640
+ "step": 700
641
+ },
642
+ {
643
+ "epoch": 0.21947449768160743,
644
+ "grad_norm": 1.4391794204711914,
645
+ "learning_rate": 1.7808346213292118e-05,
646
+ "loss": 3.5622,
647
+ "mean_token_accuracy": 0.37740132212638855,
648
+ "num_tokens": 2247593.0,
649
+ "step": 710
650
+ },
651
+ {
652
+ "epoch": 0.22256568778979907,
653
+ "grad_norm": 5.364869117736816,
654
+ "learning_rate": 1.77774343122102e-05,
655
+ "loss": 3.5855,
656
+ "mean_token_accuracy": 0.37872459217905996,
657
+ "num_tokens": 2277810.0,
658
+ "step": 720
659
+ },
660
+ {
661
+ "epoch": 0.22565687789799072,
662
+ "grad_norm": 1.3683693408966064,
663
+ "learning_rate": 1.7746522411128285e-05,
664
+ "loss": 3.4653,
665
+ "mean_token_accuracy": 0.38120819330215455,
666
+ "num_tokens": 2309457.0,
667
+ "step": 730
668
+ },
669
+ {
670
+ "epoch": 0.2287480680061824,
671
+ "grad_norm": 1.3637374639511108,
672
+ "learning_rate": 1.771561051004637e-05,
673
+ "loss": 3.5131,
674
+ "mean_token_accuracy": 0.3864184685051441,
675
+ "num_tokens": 2339852.0,
676
+ "step": 740
677
+ },
678
+ {
679
+ "epoch": 0.23183925811437403,
680
+ "grad_norm": 2.7388429641723633,
681
+ "learning_rate": 1.7684698608964452e-05,
682
+ "loss": 3.573,
683
+ "mean_token_accuracy": 0.3795412413775921,
684
+ "num_tokens": 2371444.0,
685
+ "step": 750
686
+ },
687
+ {
688
+ "epoch": 0.23493044822256567,
689
+ "grad_norm": 9.798398971557617,
690
+ "learning_rate": 1.7653786707882536e-05,
691
+ "loss": 3.5441,
692
+ "mean_token_accuracy": 0.38080229982733727,
693
+ "num_tokens": 2401769.0,
694
+ "step": 760
695
+ },
696
+ {
697
+ "epoch": 0.23802163833075735,
698
+ "grad_norm": 2.1416878700256348,
699
+ "learning_rate": 1.762287480680062e-05,
700
+ "loss": 3.5043,
701
+ "mean_token_accuracy": 0.3838089659810066,
702
+ "num_tokens": 2432410.0,
703
+ "step": 770
704
+ },
705
+ {
706
+ "epoch": 0.241112828438949,
707
+ "grad_norm": 2.211545467376709,
708
+ "learning_rate": 1.7591962905718703e-05,
709
+ "loss": 3.5341,
710
+ "mean_token_accuracy": 0.388753118366003,
711
+ "num_tokens": 2462641.0,
712
+ "step": 780
713
+ },
714
+ {
715
+ "epoch": 0.24420401854714066,
716
+ "grad_norm": 5.351387023925781,
717
+ "learning_rate": 1.7561051004636787e-05,
718
+ "loss": 3.5031,
719
+ "mean_token_accuracy": 0.3821559719741344,
720
+ "num_tokens": 2496459.0,
721
+ "step": 790
722
+ },
723
+ {
724
+ "epoch": 0.2472952086553323,
725
+ "grad_norm": 2.3877508640289307,
726
+ "learning_rate": 1.753013910355487e-05,
727
+ "loss": 3.5589,
728
+ "mean_token_accuracy": 0.3789012677967548,
729
+ "num_tokens": 2526708.0,
730
+ "step": 800
731
+ },
732
+ {
733
+ "epoch": 0.250386398763524,
734
+ "grad_norm": 1.2962738275527954,
735
+ "learning_rate": 1.7499227202472954e-05,
736
+ "loss": 3.505,
737
+ "mean_token_accuracy": 0.3823570780456066,
738
+ "num_tokens": 2561229.0,
739
+ "step": 810
740
+ },
741
+ {
742
+ "epoch": 0.2534775888717156,
743
+ "grad_norm": 1.8147461414337158,
744
+ "learning_rate": 1.7468315301391037e-05,
745
+ "loss": 3.4861,
746
+ "mean_token_accuracy": 0.3802278622984886,
747
+ "num_tokens": 2592914.0,
748
+ "step": 820
749
+ },
750
+ {
751
+ "epoch": 0.25656877897990726,
752
+ "grad_norm": 3.4690439701080322,
753
+ "learning_rate": 1.743740340030912e-05,
754
+ "loss": 3.5202,
755
+ "mean_token_accuracy": 0.38112854287028314,
756
+ "num_tokens": 2627185.0,
757
+ "step": 830
758
+ },
759
+ {
760
+ "epoch": 0.2596599690880989,
761
+ "grad_norm": 4.353795051574707,
762
+ "learning_rate": 1.7406491499227205e-05,
763
+ "loss": 3.5667,
764
+ "mean_token_accuracy": 0.37200469225645066,
765
+ "num_tokens": 2660839.0,
766
+ "step": 840
767
+ },
768
+ {
769
+ "epoch": 0.26275115919629055,
770
+ "grad_norm": 2.6319100856781006,
771
+ "learning_rate": 1.7375579598145288e-05,
772
+ "loss": 3.464,
773
+ "mean_token_accuracy": 0.38627258986234664,
774
+ "num_tokens": 2690042.0,
775
+ "step": 850
776
+ },
777
+ {
778
+ "epoch": 0.26584234930448225,
779
+ "grad_norm": 1.6525287628173828,
780
+ "learning_rate": 1.7344667697063372e-05,
781
+ "loss": 3.5616,
782
+ "mean_token_accuracy": 0.3763453342020512,
783
+ "num_tokens": 2724354.0,
784
+ "step": 860
785
+ },
786
+ {
787
+ "epoch": 0.2689335394126739,
788
+ "grad_norm": 1.9737221002578735,
789
+ "learning_rate": 1.7313755795981455e-05,
790
+ "loss": 3.5147,
791
+ "mean_token_accuracy": 0.38340551406145096,
792
+ "num_tokens": 2755591.0,
793
+ "step": 870
794
+ },
795
+ {
796
+ "epoch": 0.27202472952086554,
797
+ "grad_norm": 2.5324320793151855,
798
+ "learning_rate": 1.7282843894899536e-05,
799
+ "loss": 3.4227,
800
+ "mean_token_accuracy": 0.3909419260919094,
801
+ "num_tokens": 2786566.0,
802
+ "step": 880
803
+ },
804
+ {
805
+ "epoch": 0.2751159196290572,
806
+ "grad_norm": 3.3844659328460693,
807
+ "learning_rate": 1.7251931993817623e-05,
808
+ "loss": 3.5561,
809
+ "mean_token_accuracy": 0.37579271346330645,
810
+ "num_tokens": 2814171.0,
811
+ "step": 890
812
+ },
813
+ {
814
+ "epoch": 0.2782071097372488,
815
+ "grad_norm": 2.7620153427124023,
816
+ "learning_rate": 1.7221020092735706e-05,
817
+ "loss": 3.5042,
818
+ "mean_token_accuracy": 0.3822973191738129,
819
+ "num_tokens": 2844402.0,
820
+ "step": 900
821
+ },
822
+ {
823
+ "epoch": 0.28129829984544047,
824
+ "grad_norm": 2.089118719100952,
825
+ "learning_rate": 1.7190108191653786e-05,
826
+ "loss": 3.4583,
827
+ "mean_token_accuracy": 0.39270223304629326,
828
+ "num_tokens": 2873728.0,
829
+ "step": 910
830
+ },
831
+ {
832
+ "epoch": 0.28438948995363217,
833
+ "grad_norm": 2.496480941772461,
834
+ "learning_rate": 1.715919629057187e-05,
835
+ "loss": 3.4466,
836
+ "mean_token_accuracy": 0.3897694177925587,
837
+ "num_tokens": 2905000.0,
838
+ "step": 920
839
+ },
840
+ {
841
+ "epoch": 0.2874806800618238,
842
+ "grad_norm": 2.162785053253174,
843
+ "learning_rate": 1.7128284389489957e-05,
844
+ "loss": 3.4075,
845
+ "mean_token_accuracy": 0.39374888986349105,
846
+ "num_tokens": 2935706.0,
847
+ "step": 930
848
+ },
849
+ {
850
+ "epoch": 0.29057187017001546,
851
+ "grad_norm": 1.5756185054779053,
852
+ "learning_rate": 1.709737248840804e-05,
853
+ "loss": 3.4396,
854
+ "mean_token_accuracy": 0.38780966177582743,
855
+ "num_tokens": 2966477.0,
856
+ "step": 940
857
+ },
858
+ {
859
+ "epoch": 0.2936630602782071,
860
+ "grad_norm": 5.708625316619873,
861
+ "learning_rate": 1.706646058732612e-05,
862
+ "loss": 3.4185,
863
+ "mean_token_accuracy": 0.39326486811041833,
864
+ "num_tokens": 2996439.0,
865
+ "step": 950
866
+ },
867
+ {
868
+ "epoch": 0.29675425038639874,
869
+ "grad_norm": 2.8689119815826416,
870
+ "learning_rate": 1.7035548686244204e-05,
871
+ "loss": 3.468,
872
+ "mean_token_accuracy": 0.3818407289683819,
873
+ "num_tokens": 3026332.0,
874
+ "step": 960
875
+ },
876
+ {
877
+ "epoch": 0.29984544049459044,
878
+ "grad_norm": 4.020405292510986,
879
+ "learning_rate": 1.700463678516229e-05,
880
+ "loss": 3.4943,
881
+ "mean_token_accuracy": 0.3888055384159088,
882
+ "num_tokens": 3057606.0,
883
+ "step": 970
884
+ },
885
+ {
886
+ "epoch": 0.3029366306027821,
887
+ "grad_norm": 1.8991873264312744,
888
+ "learning_rate": 1.697372488408037e-05,
889
+ "loss": 3.4939,
890
+ "mean_token_accuracy": 0.38414665684103966,
891
+ "num_tokens": 3089449.0,
892
+ "step": 980
893
+ },
894
+ {
895
+ "epoch": 0.30602782071097373,
896
+ "grad_norm": 13.78397274017334,
897
+ "learning_rate": 1.6942812982998455e-05,
898
+ "loss": 3.5487,
899
+ "mean_token_accuracy": 0.3769320294260979,
900
+ "num_tokens": 3119537.0,
901
+ "step": 990
902
+ },
903
+ {
904
+ "epoch": 0.3091190108191654,
905
+ "grad_norm": 1.599820613861084,
906
+ "learning_rate": 1.691190108191654e-05,
907
+ "loss": 3.465,
908
+ "mean_token_accuracy": 0.3884730890393257,
909
+ "num_tokens": 3151905.0,
910
+ "step": 1000
911
+ },
912
+ {
913
+ "epoch": 0.312210200927357,
914
+ "grad_norm": 6.015178680419922,
915
+ "learning_rate": 1.6880989180834622e-05,
916
+ "loss": 3.5043,
917
+ "mean_token_accuracy": 0.38035417571663854,
918
+ "num_tokens": 3184277.0,
919
+ "step": 1010
920
+ },
921
+ {
922
+ "epoch": 0.31530139103554866,
923
+ "grad_norm": 1.854093313217163,
924
+ "learning_rate": 1.6850077279752706e-05,
925
+ "loss": 3.4233,
926
+ "mean_token_accuracy": 0.389019088447094,
927
+ "num_tokens": 3217455.0,
928
+ "step": 1020
929
+ },
930
+ {
931
+ "epoch": 0.31839258114374036,
932
+ "grad_norm": 3.024531126022339,
933
+ "learning_rate": 1.681916537867079e-05,
934
+ "loss": 3.4768,
935
+ "mean_token_accuracy": 0.39151332527399063,
936
+ "num_tokens": 3251326.0,
937
+ "step": 1030
938
+ },
939
+ {
940
+ "epoch": 0.321483771251932,
941
+ "grad_norm": 1.8012628555297852,
942
+ "learning_rate": 1.6788253477588873e-05,
943
+ "loss": 3.4955,
944
+ "mean_token_accuracy": 0.3855120025575161,
945
+ "num_tokens": 3283488.0,
946
+ "step": 1040
947
+ },
948
+ {
949
+ "epoch": 0.32457496136012365,
950
+ "grad_norm": 1.4446407556533813,
951
+ "learning_rate": 1.6757341576506957e-05,
952
+ "loss": 3.4023,
953
+ "mean_token_accuracy": 0.3883494645357132,
954
+ "num_tokens": 3315970.0,
955
+ "step": 1050
956
+ },
957
+ {
958
+ "epoch": 0.3276661514683153,
959
+ "grad_norm": 2.132194995880127,
960
+ "learning_rate": 1.672642967542504e-05,
961
+ "loss": 3.5195,
962
+ "mean_token_accuracy": 0.38440208286046984,
963
+ "num_tokens": 3348645.0,
964
+ "step": 1060
965
+ },
966
+ {
967
+ "epoch": 0.33075734157650694,
968
+ "grad_norm": 2.436528444290161,
969
+ "learning_rate": 1.6695517774343124e-05,
970
+ "loss": 3.4595,
971
+ "mean_token_accuracy": 0.3885770753026009,
972
+ "num_tokens": 3378707.0,
973
+ "step": 1070
974
+ },
975
+ {
976
+ "epoch": 0.33384853168469864,
977
+ "grad_norm": 2.0806350708007812,
978
+ "learning_rate": 1.6664605873261207e-05,
979
+ "loss": 3.4576,
980
+ "mean_token_accuracy": 0.38848345205187795,
981
+ "num_tokens": 3410231.0,
982
+ "step": 1080
983
+ },
984
+ {
985
+ "epoch": 0.3369397217928903,
986
+ "grad_norm": 2.142319679260254,
987
+ "learning_rate": 1.663369397217929e-05,
988
+ "loss": 3.4644,
989
+ "mean_token_accuracy": 0.3861194223165512,
990
+ "num_tokens": 3444065.0,
991
+ "step": 1090
992
+ },
993
+ {
994
+ "epoch": 0.3400309119010819,
995
+ "grad_norm": 2.380552053451538,
996
+ "learning_rate": 1.6602782071097374e-05,
997
+ "loss": 3.3916,
998
+ "mean_token_accuracy": 0.39232398346066477,
999
+ "num_tokens": 3474974.0,
1000
+ "step": 1100
1001
+ },
1002
+ {
1003
+ "epoch": 0.34312210200927357,
1004
+ "grad_norm": 1.6462984085083008,
1005
+ "learning_rate": 1.6571870170015458e-05,
1006
+ "loss": 3.3687,
1007
+ "mean_token_accuracy": 0.3944365203380585,
1008
+ "num_tokens": 3507101.0,
1009
+ "step": 1110
1010
+ },
1011
+ {
1012
+ "epoch": 0.3462132921174652,
1013
+ "grad_norm": 4.342376232147217,
1014
+ "learning_rate": 1.654095826893354e-05,
1015
+ "loss": 3.4589,
1016
+ "mean_token_accuracy": 0.3856883034110069,
1017
+ "num_tokens": 3538893.0,
1018
+ "step": 1120
1019
+ },
1020
+ {
1021
+ "epoch": 0.34930448222565685,
1022
+ "grad_norm": 1.8314056396484375,
1023
+ "learning_rate": 1.6510046367851625e-05,
1024
+ "loss": 3.6277,
1025
+ "mean_token_accuracy": 0.3779118649661541,
1026
+ "num_tokens": 3567776.0,
1027
+ "step": 1130
1028
+ },
1029
+ {
1030
+ "epoch": 0.35239567233384855,
1031
+ "grad_norm": 1.7731289863586426,
1032
+ "learning_rate": 1.647913446676971e-05,
1033
+ "loss": 3.4375,
1034
+ "mean_token_accuracy": 0.38362068235874175,
1035
+ "num_tokens": 3600096.0,
1036
+ "step": 1140
1037
+ },
1038
+ {
1039
+ "epoch": 0.3554868624420402,
1040
+ "grad_norm": 1.4724918603897095,
1041
+ "learning_rate": 1.644822256568779e-05,
1042
+ "loss": 3.4787,
1043
+ "mean_token_accuracy": 0.39140490964055064,
1044
+ "num_tokens": 3629645.0,
1045
+ "step": 1150
1046
+ },
1047
+ {
1048
+ "epoch": 0.35857805255023184,
1049
+ "grad_norm": 2.0459768772125244,
1050
+ "learning_rate": 1.6417310664605876e-05,
1051
+ "loss": 3.4346,
1052
+ "mean_token_accuracy": 0.38677491843700407,
1053
+ "num_tokens": 3663559.0,
1054
+ "step": 1160
1055
+ },
1056
+ {
1057
+ "epoch": 0.3616692426584235,
1058
+ "grad_norm": 3.235039710998535,
1059
+ "learning_rate": 1.638639876352396e-05,
1060
+ "loss": 3.4915,
1061
+ "mean_token_accuracy": 0.38628031834959986,
1062
+ "num_tokens": 3695457.0,
1063
+ "step": 1170
1064
+ },
1065
+ {
1066
+ "epoch": 0.36476043276661513,
1067
+ "grad_norm": 1.2474194765090942,
1068
+ "learning_rate": 1.635548686244204e-05,
1069
+ "loss": 3.3999,
1070
+ "mean_token_accuracy": 0.39414023533463477,
1071
+ "num_tokens": 3725294.0,
1072
+ "step": 1180
1073
+ },
1074
+ {
1075
+ "epoch": 0.3678516228748068,
1076
+ "grad_norm": 1.3599259853363037,
1077
+ "learning_rate": 1.6324574961360123e-05,
1078
+ "loss": 3.3712,
1079
+ "mean_token_accuracy": 0.39129213988780975,
1080
+ "num_tokens": 3756722.0,
1081
+ "step": 1190
1082
+ },
1083
+ {
1084
+ "epoch": 0.37094281298299847,
1085
+ "grad_norm": 1.4785717725753784,
1086
+ "learning_rate": 1.629366306027821e-05,
1087
+ "loss": 3.3837,
1088
+ "mean_token_accuracy": 0.39773035794496536,
1089
+ "num_tokens": 3786396.0,
1090
+ "step": 1200
1091
+ },
1092
+ {
1093
+ "epoch": 0.3740340030911901,
1094
+ "grad_norm": 1.1180949211120605,
1095
+ "learning_rate": 1.6262751159196294e-05,
1096
+ "loss": 3.374,
1097
+ "mean_token_accuracy": 0.3942295677959919,
1098
+ "num_tokens": 3816940.0,
1099
+ "step": 1210
1100
+ },
1101
+ {
1102
+ "epoch": 0.37712519319938176,
1103
+ "grad_norm": 3.58443546295166,
1104
+ "learning_rate": 1.6231839258114374e-05,
1105
+ "loss": 3.3892,
1106
+ "mean_token_accuracy": 0.3918235659599304,
1107
+ "num_tokens": 3845728.0,
1108
+ "step": 1220
1109
+ },
1110
+ {
1111
+ "epoch": 0.3802163833075734,
1112
+ "grad_norm": 7.910126686096191,
1113
+ "learning_rate": 1.6200927357032458e-05,
1114
+ "loss": 3.4141,
1115
+ "mean_token_accuracy": 0.39633639603853227,
1116
+ "num_tokens": 3876251.0,
1117
+ "step": 1230
1118
+ },
1119
+ {
1120
+ "epoch": 0.38330757341576505,
1121
+ "grad_norm": 1.4737247228622437,
1122
+ "learning_rate": 1.617001545595054e-05,
1123
+ "loss": 3.3891,
1124
+ "mean_token_accuracy": 0.38747691363096237,
1125
+ "num_tokens": 3907979.0,
1126
+ "step": 1240
1127
+ },
1128
+ {
1129
+ "epoch": 0.38639876352395675,
1130
+ "grad_norm": 2.543823003768921,
1131
+ "learning_rate": 1.6139103554868625e-05,
1132
+ "loss": 3.3857,
1133
+ "mean_token_accuracy": 0.3897668160498142,
1134
+ "num_tokens": 3939556.0,
1135
+ "step": 1250
1136
+ },
1137
+ {
1138
+ "epoch": 0.3894899536321484,
1139
+ "grad_norm": 1.8837332725524902,
1140
+ "learning_rate": 1.610819165378671e-05,
1141
+ "loss": 3.3795,
1142
+ "mean_token_accuracy": 0.39769657924771307,
1143
+ "num_tokens": 3970481.0,
1144
+ "step": 1260
1145
+ },
1146
+ {
1147
+ "epoch": 0.39258114374034003,
1148
+ "grad_norm": 3.573788642883301,
1149
+ "learning_rate": 1.6077279752704792e-05,
1150
+ "loss": 3.3105,
1151
+ "mean_token_accuracy": 0.4041719429194927,
1152
+ "num_tokens": 3999666.0,
1153
+ "step": 1270
1154
+ },
1155
+ {
1156
+ "epoch": 0.3956723338485317,
1157
+ "grad_norm": 1.7824413776397705,
1158
+ "learning_rate": 1.6046367851622876e-05,
1159
+ "loss": 3.4468,
1160
+ "mean_token_accuracy": 0.3911862142384052,
1161
+ "num_tokens": 4031876.0,
1162
+ "step": 1280
1163
+ },
1164
+ {
1165
+ "epoch": 0.3987635239567233,
1166
+ "grad_norm": 1.6730329990386963,
1167
+ "learning_rate": 1.601545595054096e-05,
1168
+ "loss": 3.4193,
1169
+ "mean_token_accuracy": 0.39065720662474634,
1170
+ "num_tokens": 4064132.0,
1171
+ "step": 1290
1172
+ },
1173
+ {
1174
+ "epoch": 0.401854714064915,
1175
+ "grad_norm": 1.421411395072937,
1176
+ "learning_rate": 1.5984544049459043e-05,
1177
+ "loss": 3.4292,
1178
+ "mean_token_accuracy": 0.3871223643422127,
1179
+ "num_tokens": 4095074.0,
1180
+ "step": 1300
1181
+ },
1182
+ {
1183
+ "epoch": 0.40494590417310666,
1184
+ "grad_norm": 1.7248343229293823,
1185
+ "learning_rate": 1.5953632148377126e-05,
1186
+ "loss": 3.379,
1187
+ "mean_token_accuracy": 0.39316892698407174,
1188
+ "num_tokens": 4130858.0,
1189
+ "step": 1310
1190
+ },
1191
+ {
1192
+ "epoch": 0.4080370942812983,
1193
+ "grad_norm": 2.6556711196899414,
1194
+ "learning_rate": 1.592272024729521e-05,
1195
+ "loss": 3.4088,
1196
+ "mean_token_accuracy": 0.38396543338894845,
1197
+ "num_tokens": 4163277.0,
1198
+ "step": 1320
1199
+ },
1200
+ {
1201
+ "epoch": 0.41112828438948995,
1202
+ "grad_norm": 2.2201597690582275,
1203
+ "learning_rate": 1.5891808346213294e-05,
1204
+ "loss": 3.446,
1205
+ "mean_token_accuracy": 0.38794904500246047,
1206
+ "num_tokens": 4195806.0,
1207
+ "step": 1330
1208
+ },
1209
+ {
1210
+ "epoch": 0.4142194744976816,
1211
+ "grad_norm": 1.4001938104629517,
1212
+ "learning_rate": 1.5860896445131377e-05,
1213
+ "loss": 3.4074,
1214
+ "mean_token_accuracy": 0.3959184519946575,
1215
+ "num_tokens": 4225811.0,
1216
+ "step": 1340
1217
+ },
1218
+ {
1219
+ "epoch": 0.41731066460587324,
1220
+ "grad_norm": 1.9553899765014648,
1221
+ "learning_rate": 1.582998454404946e-05,
1222
+ "loss": 3.353,
1223
+ "mean_token_accuracy": 0.392233844101429,
1224
+ "num_tokens": 4259418.0,
1225
+ "step": 1350
1226
+ },
1227
+ {
1228
+ "epoch": 0.42040185471406494,
1229
+ "grad_norm": 2.1650876998901367,
1230
+ "learning_rate": 1.5799072642967544e-05,
1231
+ "loss": 3.3014,
1232
+ "mean_token_accuracy": 0.40291827023029325,
1233
+ "num_tokens": 4292593.0,
1234
+ "step": 1360
1235
+ },
1236
+ {
1237
+ "epoch": 0.4234930448222566,
1238
+ "grad_norm": 1.6802010536193848,
1239
+ "learning_rate": 1.5768160741885628e-05,
1240
+ "loss": 3.3615,
1241
+ "mean_token_accuracy": 0.3953541323542595,
1242
+ "num_tokens": 4325628.0,
1243
+ "step": 1370
1244
+ },
1245
+ {
1246
+ "epoch": 0.4265842349304482,
1247
+ "grad_norm": 2.8798279762268066,
1248
+ "learning_rate": 1.573724884080371e-05,
1249
+ "loss": 3.4365,
1250
+ "mean_token_accuracy": 0.39370308369398116,
1251
+ "num_tokens": 4358459.0,
1252
+ "step": 1380
1253
+ },
1254
+ {
1255
+ "epoch": 0.42967542503863987,
1256
+ "grad_norm": 1.523694634437561,
1257
+ "learning_rate": 1.5706336939721795e-05,
1258
+ "loss": 3.3153,
1259
+ "mean_token_accuracy": 0.3944700941443443,
1260
+ "num_tokens": 4390529.0,
1261
+ "step": 1390
1262
+ },
1263
+ {
1264
+ "epoch": 0.4327666151468315,
1265
+ "grad_norm": 2.0799732208251953,
1266
+ "learning_rate": 1.567542503863988e-05,
1267
+ "loss": 3.4242,
1268
+ "mean_token_accuracy": 0.389276672154665,
1269
+ "num_tokens": 4424045.0,
1270
+ "step": 1400
1271
+ },
1272
+ {
1273
+ "epoch": 0.43585780525502316,
1274
+ "grad_norm": 1.3916538953781128,
1275
+ "learning_rate": 1.5644513137557962e-05,
1276
+ "loss": 3.444,
1277
+ "mean_token_accuracy": 0.388773063570261,
1278
+ "num_tokens": 4456115.0,
1279
+ "step": 1410
1280
+ },
1281
+ {
1282
+ "epoch": 0.43894899536321486,
1283
+ "grad_norm": 11.072097778320312,
1284
+ "learning_rate": 1.5613601236476042e-05,
1285
+ "loss": 3.2926,
1286
+ "mean_token_accuracy": 0.40686351582407954,
1287
+ "num_tokens": 4490990.0,
1288
+ "step": 1420
1289
+ },
1290
+ {
1291
+ "epoch": 0.4420401854714065,
1292
+ "grad_norm": 1.8008073568344116,
1293
+ "learning_rate": 1.558268933539413e-05,
1294
+ "loss": 3.4062,
1295
+ "mean_token_accuracy": 0.3938455879688263,
1296
+ "num_tokens": 4519847.0,
1297
+ "step": 1430
1298
+ },
1299
+ {
1300
+ "epoch": 0.44513137557959814,
1301
+ "grad_norm": 4.065845012664795,
1302
+ "learning_rate": 1.5551777434312213e-05,
1303
+ "loss": 3.3776,
1304
+ "mean_token_accuracy": 0.4004744917154312,
1305
+ "num_tokens": 4551679.0,
1306
+ "step": 1440
1307
+ },
1308
+ {
1309
+ "epoch": 0.4482225656877898,
1310
+ "grad_norm": 2.4614624977111816,
1311
+ "learning_rate": 1.5520865533230297e-05,
1312
+ "loss": 3.324,
1313
+ "mean_token_accuracy": 0.4031891174614429,
1314
+ "num_tokens": 4582622.0,
1315
+ "step": 1450
1316
+ },
1317
+ {
1318
+ "epoch": 0.45131375579598143,
1319
+ "grad_norm": 2.6623904705047607,
1320
+ "learning_rate": 1.5489953632148377e-05,
1321
+ "loss": 3.414,
1322
+ "mean_token_accuracy": 0.39484291821718215,
1323
+ "num_tokens": 4613478.0,
1324
+ "step": 1460
1325
+ },
1326
+ {
1327
+ "epoch": 0.45440494590417313,
1328
+ "grad_norm": 2.342698574066162,
1329
+ "learning_rate": 1.545904173106646e-05,
1330
+ "loss": 3.4348,
1331
+ "mean_token_accuracy": 0.38681296780705454,
1332
+ "num_tokens": 4647495.0,
1333
+ "step": 1470
1334
+ },
1335
+ {
1336
+ "epoch": 0.4574961360123648,
1337
+ "grad_norm": 2.6113431453704834,
1338
+ "learning_rate": 1.5428129829984547e-05,
1339
+ "loss": 3.39,
1340
+ "mean_token_accuracy": 0.391095020622015,
1341
+ "num_tokens": 4681809.0,
1342
+ "step": 1480
1343
+ },
1344
+ {
1345
+ "epoch": 0.4605873261205564,
1346
+ "grad_norm": 1.3145941495895386,
1347
+ "learning_rate": 1.5397217928902627e-05,
1348
+ "loss": 3.3677,
1349
+ "mean_token_accuracy": 0.38954789489507674,
1350
+ "num_tokens": 4716308.0,
1351
+ "step": 1490
1352
+ },
1353
+ {
1354
+ "epoch": 0.46367851622874806,
1355
+ "grad_norm": 1.8231385946273804,
1356
+ "learning_rate": 1.536630602782071e-05,
1357
+ "loss": 3.4199,
1358
+ "mean_token_accuracy": 0.38780914843082426,
1359
+ "num_tokens": 4747835.0,
1360
+ "step": 1500
1361
+ },
1362
+ {
1363
+ "epoch": 0.4667697063369397,
1364
+ "grad_norm": 2.2594892978668213,
1365
+ "learning_rate": 1.5335394126738795e-05,
1366
+ "loss": 3.3937,
1367
+ "mean_token_accuracy": 0.3959015667438507,
1368
+ "num_tokens": 4776486.0,
1369
+ "step": 1510
1370
+ },
1371
+ {
1372
+ "epoch": 0.46986089644513135,
1373
+ "grad_norm": 3.840742349624634,
1374
+ "learning_rate": 1.5304482225656878e-05,
1375
+ "loss": 3.3087,
1376
+ "mean_token_accuracy": 0.407059845328331,
1377
+ "num_tokens": 4805943.0,
1378
+ "step": 1520
1379
+ },
1380
+ {
1381
+ "epoch": 0.47295208655332305,
1382
+ "grad_norm": 4.683995246887207,
1383
+ "learning_rate": 1.5273570324574962e-05,
1384
+ "loss": 3.3373,
1385
+ "mean_token_accuracy": 0.39871800169348715,
1386
+ "num_tokens": 4838440.0,
1387
+ "step": 1530
1388
+ },
1389
+ {
1390
+ "epoch": 0.4760432766615147,
1391
+ "grad_norm": 1.2700860500335693,
1392
+ "learning_rate": 1.5242658423493047e-05,
1393
+ "loss": 3.3339,
1394
+ "mean_token_accuracy": 0.39955019652843476,
1395
+ "num_tokens": 4870809.0,
1396
+ "step": 1540
1397
+ },
1398
+ {
1399
+ "epoch": 0.47913446676970634,
1400
+ "grad_norm": 2.15336537361145,
1401
+ "learning_rate": 1.521174652241113e-05,
1402
+ "loss": 3.3699,
1403
+ "mean_token_accuracy": 0.39904908165335656,
1404
+ "num_tokens": 4903332.0,
1405
+ "step": 1550
1406
+ },
1407
+ {
1408
+ "epoch": 0.482225656877898,
1409
+ "grad_norm": 1.299379587173462,
1410
+ "learning_rate": 1.5180834621329213e-05,
1411
+ "loss": 3.4024,
1412
+ "mean_token_accuracy": 0.3913417667150497,
1413
+ "num_tokens": 4933852.0,
1414
+ "step": 1560
1415
+ },
1416
+ {
1417
+ "epoch": 0.4853168469860896,
1418
+ "grad_norm": 1.1829091310501099,
1419
+ "learning_rate": 1.5149922720247296e-05,
1420
+ "loss": 3.3253,
1421
+ "mean_token_accuracy": 0.39646707102656364,
1422
+ "num_tokens": 4966530.0,
1423
+ "step": 1570
1424
+ },
1425
+ {
1426
+ "epoch": 0.4884080370942813,
1427
+ "grad_norm": 1.2553237676620483,
1428
+ "learning_rate": 1.511901081916538e-05,
1429
+ "loss": 3.2771,
1430
+ "mean_token_accuracy": 0.405765625834465,
1431
+ "num_tokens": 4998456.0,
1432
+ "step": 1580
1433
+ },
1434
+ {
1435
+ "epoch": 0.49149922720247297,
1436
+ "grad_norm": 2.037930488586426,
1437
+ "learning_rate": 1.5088098918083462e-05,
1438
+ "loss": 3.3754,
1439
+ "mean_token_accuracy": 0.3952351205050945,
1440
+ "num_tokens": 5028995.0,
1441
+ "step": 1590
1442
+ },
1443
+ {
1444
+ "epoch": 0.4945904173106646,
1445
+ "grad_norm": 1.090571403503418,
1446
+ "learning_rate": 1.5057187017001547e-05,
1447
+ "loss": 3.2832,
1448
+ "mean_token_accuracy": 0.4088763400912285,
1449
+ "num_tokens": 5057407.0,
1450
+ "step": 1600
1451
+ },
1452
+ {
1453
+ "epoch": 0.49768160741885625,
1454
+ "grad_norm": 2.4721672534942627,
1455
+ "learning_rate": 1.502627511591963e-05,
1456
+ "loss": 3.3535,
1457
+ "mean_token_accuracy": 0.3982778422534466,
1458
+ "num_tokens": 5088330.0,
1459
+ "step": 1610
1460
+ },
1461
+ {
1462
+ "epoch": 0.500772797527048,
1463
+ "grad_norm": 1.3971518278121948,
1464
+ "learning_rate": 1.4995363214837714e-05,
1465
+ "loss": 3.359,
1466
+ "mean_token_accuracy": 0.3978785939514637,
1467
+ "num_tokens": 5115511.0,
1468
+ "step": 1620
1469
+ },
1470
+ {
1471
+ "epoch": 0.5038639876352395,
1472
+ "grad_norm": 1.766021490097046,
1473
+ "learning_rate": 1.4964451313755796e-05,
1474
+ "loss": 3.3388,
1475
+ "mean_token_accuracy": 0.4023651979863644,
1476
+ "num_tokens": 5145560.0,
1477
+ "step": 1630
1478
+ },
1479
+ {
1480
+ "epoch": 0.5069551777434312,
1481
+ "grad_norm": 0.9848290681838989,
1482
+ "learning_rate": 1.4933539412673881e-05,
1483
+ "loss": 3.3312,
1484
+ "mean_token_accuracy": 0.4001339070498943,
1485
+ "num_tokens": 5178491.0,
1486
+ "step": 1640
1487
+ },
1488
+ {
1489
+ "epoch": 0.5100463678516228,
1490
+ "grad_norm": 5.008669376373291,
1491
+ "learning_rate": 1.4902627511591965e-05,
1492
+ "loss": 3.3395,
1493
+ "mean_token_accuracy": 0.39800570756196973,
1494
+ "num_tokens": 5209376.0,
1495
+ "step": 1650
1496
+ },
1497
+ {
1498
+ "epoch": 0.5131375579598145,
1499
+ "grad_norm": 3.3218369483947754,
1500
+ "learning_rate": 1.4871715610510047e-05,
1501
+ "loss": 3.398,
1502
+ "mean_token_accuracy": 0.3898357510566711,
1503
+ "num_tokens": 5240240.0,
1504
+ "step": 1660
1505
+ },
1506
+ {
1507
+ "epoch": 0.5162287480680062,
1508
+ "grad_norm": 1.3590713739395142,
1509
+ "learning_rate": 1.484080370942813e-05,
1510
+ "loss": 3.3246,
1511
+ "mean_token_accuracy": 0.39972665831446647,
1512
+ "num_tokens": 5269658.0,
1513
+ "step": 1670
1514
+ },
1515
+ {
1516
+ "epoch": 0.5193199381761978,
1517
+ "grad_norm": 1.647360920906067,
1518
+ "learning_rate": 1.4809891808346216e-05,
1519
+ "loss": 3.3101,
1520
+ "mean_token_accuracy": 0.4039941616356373,
1521
+ "num_tokens": 5301959.0,
1522
+ "step": 1680
1523
+ },
1524
+ {
1525
+ "epoch": 0.5224111282843895,
1526
+ "grad_norm": 1.3231589794158936,
1527
+ "learning_rate": 1.4778979907264298e-05,
1528
+ "loss": 3.2715,
1529
+ "mean_token_accuracy": 0.409288527816534,
1530
+ "num_tokens": 5332939.0,
1531
+ "step": 1690
1532
+ },
1533
+ {
1534
+ "epoch": 0.5255023183925811,
1535
+ "grad_norm": 1.8494716882705688,
1536
+ "learning_rate": 1.4748068006182381e-05,
1537
+ "loss": 3.3472,
1538
+ "mean_token_accuracy": 0.4054514840245247,
1539
+ "num_tokens": 5361660.0,
1540
+ "step": 1700
1541
+ },
1542
+ {
1543
+ "epoch": 0.5285935085007728,
1544
+ "grad_norm": 1.1796019077301025,
1545
+ "learning_rate": 1.4717156105100465e-05,
1546
+ "loss": 3.3374,
1547
+ "mean_token_accuracy": 0.40053225085139277,
1548
+ "num_tokens": 5390818.0,
1549
+ "step": 1710
1550
+ },
1551
+ {
1552
+ "epoch": 0.5316846986089645,
1553
+ "grad_norm": 1.4589906930923462,
1554
+ "learning_rate": 1.468624420401855e-05,
1555
+ "loss": 3.3532,
1556
+ "mean_token_accuracy": 0.3908236466348171,
1557
+ "num_tokens": 5423948.0,
1558
+ "step": 1720
1559
+ },
1560
+ {
1561
+ "epoch": 0.5347758887171561,
1562
+ "grad_norm": 2.48760986328125,
1563
+ "learning_rate": 1.4655332302936632e-05,
1564
+ "loss": 3.3916,
1565
+ "mean_token_accuracy": 0.39238951057195665,
1566
+ "num_tokens": 5456721.0,
1567
+ "step": 1730
1568
+ },
1569
+ {
1570
+ "epoch": 0.5378670788253478,
1571
+ "grad_norm": 1.3151116371154785,
1572
+ "learning_rate": 1.4624420401854715e-05,
1573
+ "loss": 3.3755,
1574
+ "mean_token_accuracy": 0.39836090207099917,
1575
+ "num_tokens": 5485487.0,
1576
+ "step": 1740
1577
+ },
1578
+ {
1579
+ "epoch": 0.5409582689335394,
1580
+ "grad_norm": 1.3862218856811523,
1581
+ "learning_rate": 1.4593508500772799e-05,
1582
+ "loss": 3.3972,
1583
+ "mean_token_accuracy": 0.39552291929721833,
1584
+ "num_tokens": 5517812.0,
1585
+ "step": 1750
1586
+ },
1587
+ {
1588
+ "epoch": 0.5440494590417311,
1589
+ "grad_norm": 1.4198209047317505,
1590
+ "learning_rate": 1.4562596599690881e-05,
1591
+ "loss": 3.323,
1592
+ "mean_token_accuracy": 0.4038604758679867,
1593
+ "num_tokens": 5549085.0,
1594
+ "step": 1760
1595
+ },
1596
+ {
1597
+ "epoch": 0.5471406491499228,
1598
+ "grad_norm": 2.063263416290283,
1599
+ "learning_rate": 1.4531684698608966e-05,
1600
+ "loss": 3.2286,
1601
+ "mean_token_accuracy": 0.40857273861765864,
1602
+ "num_tokens": 5580826.0,
1603
+ "step": 1770
1604
+ },
1605
+ {
1606
+ "epoch": 0.5502318392581144,
1607
+ "grad_norm": 1.5246375799179077,
1608
+ "learning_rate": 1.450077279752705e-05,
1609
+ "loss": 3.3077,
1610
+ "mean_token_accuracy": 0.3984913781285286,
1611
+ "num_tokens": 5611654.0,
1612
+ "step": 1780
1613
+ },
1614
+ {
1615
+ "epoch": 0.5533230293663061,
1616
+ "grad_norm": 3.1407299041748047,
1617
+ "learning_rate": 1.4469860896445132e-05,
1618
+ "loss": 3.3631,
1619
+ "mean_token_accuracy": 0.4005904957652092,
1620
+ "num_tokens": 5643972.0,
1621
+ "step": 1790
1622
+ },
1623
+ {
1624
+ "epoch": 0.5564142194744977,
1625
+ "grad_norm": 1.1959761381149292,
1626
+ "learning_rate": 1.4438948995363215e-05,
1627
+ "loss": 3.3863,
1628
+ "mean_token_accuracy": 0.39492699652910235,
1629
+ "num_tokens": 5674824.0,
1630
+ "step": 1800
1631
+ },
1632
+ {
1633
+ "epoch": 0.5595054095826894,
1634
+ "grad_norm": 1.6108070611953735,
1635
+ "learning_rate": 1.44080370942813e-05,
1636
+ "loss": 3.3524,
1637
+ "mean_token_accuracy": 0.3943831264972687,
1638
+ "num_tokens": 5710374.0,
1639
+ "step": 1810
1640
+ },
1641
+ {
1642
+ "epoch": 0.5625965996908809,
1643
+ "grad_norm": 2.3074188232421875,
1644
+ "learning_rate": 1.4377125193199384e-05,
1645
+ "loss": 3.4149,
1646
+ "mean_token_accuracy": 0.3902184680104256,
1647
+ "num_tokens": 5740450.0,
1648
+ "step": 1820
1649
+ },
1650
+ {
1651
+ "epoch": 0.5656877897990726,
1652
+ "grad_norm": 1.2210988998413086,
1653
+ "learning_rate": 1.4346213292117466e-05,
1654
+ "loss": 3.368,
1655
+ "mean_token_accuracy": 0.3951147675514221,
1656
+ "num_tokens": 5769865.0,
1657
+ "step": 1830
1658
+ },
1659
+ {
1660
+ "epoch": 0.5687789799072643,
1661
+ "grad_norm": 1.301492691040039,
1662
+ "learning_rate": 1.431530139103555e-05,
1663
+ "loss": 3.4011,
1664
+ "mean_token_accuracy": 0.39827719777822496,
1665
+ "num_tokens": 5798647.0,
1666
+ "step": 1840
1667
+ },
1668
+ {
1669
+ "epoch": 0.5718701700154559,
1670
+ "grad_norm": 1.8520913124084473,
1671
+ "learning_rate": 1.4284389489953633e-05,
1672
+ "loss": 3.3297,
1673
+ "mean_token_accuracy": 0.401262603700161,
1674
+ "num_tokens": 5830176.0,
1675
+ "step": 1850
1676
+ },
1677
+ {
1678
+ "epoch": 0.5749613601236476,
1679
+ "grad_norm": 1.7829524278640747,
1680
+ "learning_rate": 1.4253477588871715e-05,
1681
+ "loss": 3.2177,
1682
+ "mean_token_accuracy": 0.4113538973033428,
1683
+ "num_tokens": 5860769.0,
1684
+ "step": 1860
1685
+ },
1686
+ {
1687
+ "epoch": 0.5780525502318392,
1688
+ "grad_norm": 1.5205345153808594,
1689
+ "learning_rate": 1.42225656877898e-05,
1690
+ "loss": 3.4594,
1691
+ "mean_token_accuracy": 0.3904502220451832,
1692
+ "num_tokens": 5896213.0,
1693
+ "step": 1870
1694
+ },
1695
+ {
1696
+ "epoch": 0.5811437403400309,
1697
+ "grad_norm": 1.492475986480713,
1698
+ "learning_rate": 1.4191653786707884e-05,
1699
+ "loss": 3.2639,
1700
+ "mean_token_accuracy": 0.40337538048624993,
1701
+ "num_tokens": 5928539.0,
1702
+ "step": 1880
1703
+ },
1704
+ {
1705
+ "epoch": 0.5842349304482226,
1706
+ "grad_norm": 2.07590651512146,
1707
+ "learning_rate": 1.4160741885625968e-05,
1708
+ "loss": 3.3918,
1709
+ "mean_token_accuracy": 0.39563274309039115,
1710
+ "num_tokens": 5960442.0,
1711
+ "step": 1890
1712
+ },
1713
+ {
1714
+ "epoch": 0.5873261205564142,
1715
+ "grad_norm": 2.6959567070007324,
1716
+ "learning_rate": 1.412982998454405e-05,
1717
+ "loss": 3.3494,
1718
+ "mean_token_accuracy": 0.40076613500714303,
1719
+ "num_tokens": 5990404.0,
1720
+ "step": 1900
1721
+ },
1722
+ {
1723
+ "epoch": 0.5904173106646059,
1724
+ "grad_norm": 1.291227102279663,
1725
+ "learning_rate": 1.4098918083462135e-05,
1726
+ "loss": 3.3794,
1727
+ "mean_token_accuracy": 0.3960807867348194,
1728
+ "num_tokens": 6023418.0,
1729
+ "step": 1910
1730
+ },
1731
+ {
1732
+ "epoch": 0.5935085007727975,
1733
+ "grad_norm": 3.709761381149292,
1734
+ "learning_rate": 1.4068006182380218e-05,
1735
+ "loss": 3.3441,
1736
+ "mean_token_accuracy": 0.3981980659067631,
1737
+ "num_tokens": 6051604.0,
1738
+ "step": 1920
1739
+ },
1740
+ {
1741
+ "epoch": 0.5965996908809892,
1742
+ "grad_norm": 1.5294419527053833,
1743
+ "learning_rate": 1.40370942812983e-05,
1744
+ "loss": 3.3821,
1745
+ "mean_token_accuracy": 0.40094061717391016,
1746
+ "num_tokens": 6081609.0,
1747
+ "step": 1930
1748
+ },
1749
+ {
1750
+ "epoch": 0.5996908809891809,
1751
+ "grad_norm": 1.792324185371399,
1752
+ "learning_rate": 1.4006182380216384e-05,
1753
+ "loss": 3.3228,
1754
+ "mean_token_accuracy": 0.40257288739085195,
1755
+ "num_tokens": 6113487.0,
1756
+ "step": 1940
1757
+ },
1758
+ {
1759
+ "epoch": 0.6027820710973725,
1760
+ "grad_norm": 1.3564989566802979,
1761
+ "learning_rate": 1.3975270479134469e-05,
1762
+ "loss": 3.3603,
1763
+ "mean_token_accuracy": 0.3971732698380947,
1764
+ "num_tokens": 6143546.0,
1765
+ "step": 1950
1766
+ },
1767
+ {
1768
+ "epoch": 0.6058732612055642,
1769
+ "grad_norm": 1.8166826963424683,
1770
+ "learning_rate": 1.3944358578052551e-05,
1771
+ "loss": 3.3511,
1772
+ "mean_token_accuracy": 0.3947428591549397,
1773
+ "num_tokens": 6177302.0,
1774
+ "step": 1960
1775
+ },
1776
+ {
1777
+ "epoch": 0.6089644513137558,
1778
+ "grad_norm": 2.3519530296325684,
1779
+ "learning_rate": 1.3913446676970635e-05,
1780
+ "loss": 3.3258,
1781
+ "mean_token_accuracy": 0.40014824345707894,
1782
+ "num_tokens": 6209913.0,
1783
+ "step": 1970
1784
+ },
1785
+ {
1786
+ "epoch": 0.6120556414219475,
1787
+ "grad_norm": 1.9432330131530762,
1788
+ "learning_rate": 1.3882534775888718e-05,
1789
+ "loss": 3.2683,
1790
+ "mean_token_accuracy": 0.40405927672982217,
1791
+ "num_tokens": 6241773.0,
1792
+ "step": 1980
1793
+ },
1794
+ {
1795
+ "epoch": 0.615146831530139,
1796
+ "grad_norm": 1.2190839052200317,
1797
+ "learning_rate": 1.3851622874806803e-05,
1798
+ "loss": 3.3476,
1799
+ "mean_token_accuracy": 0.4006108805537224,
1800
+ "num_tokens": 6274930.0,
1801
+ "step": 1990
1802
+ },
1803
+ {
1804
+ "epoch": 0.6182380216383307,
1805
+ "grad_norm": 1.395822525024414,
1806
+ "learning_rate": 1.3820710973724885e-05,
1807
+ "loss": 3.2706,
1808
+ "mean_token_accuracy": 0.40902155488729475,
1809
+ "num_tokens": 6306366.0,
1810
+ "step": 2000
1811
+ },
1812
+ {
1813
+ "epoch": 0.6213292117465224,
1814
+ "grad_norm": 3.3072211742401123,
1815
+ "learning_rate": 1.3789799072642969e-05,
1816
+ "loss": 3.2842,
1817
+ "mean_token_accuracy": 0.4068531468510628,
1818
+ "num_tokens": 6336620.0,
1819
+ "step": 2010
1820
+ },
1821
+ {
1822
+ "epoch": 0.624420401854714,
1823
+ "grad_norm": 6.775637626647949,
1824
+ "learning_rate": 1.3758887171561052e-05,
1825
+ "loss": 3.3359,
1826
+ "mean_token_accuracy": 0.399787887185812,
1827
+ "num_tokens": 6368783.0,
1828
+ "step": 2020
1829
+ },
1830
+ {
1831
+ "epoch": 0.6275115919629057,
1832
+ "grad_norm": 2.236809253692627,
1833
+ "learning_rate": 1.3727975270479134e-05,
1834
+ "loss": 3.1501,
1835
+ "mean_token_accuracy": 0.4164234817028046,
1836
+ "num_tokens": 6396050.0,
1837
+ "step": 2030
1838
+ },
1839
+ {
1840
+ "epoch": 0.6306027820710973,
1841
+ "grad_norm": 1.864715814590454,
1842
+ "learning_rate": 1.369706336939722e-05,
1843
+ "loss": 3.2679,
1844
+ "mean_token_accuracy": 0.40604618191719055,
1845
+ "num_tokens": 6428401.0,
1846
+ "step": 2040
1847
+ },
1848
+ {
1849
+ "epoch": 0.633693972179289,
1850
+ "grad_norm": 13.025823593139648,
1851
+ "learning_rate": 1.3666151468315303e-05,
1852
+ "loss": 3.2663,
1853
+ "mean_token_accuracy": 0.39983370155096054,
1854
+ "num_tokens": 6463617.0,
1855
+ "step": 2050
1856
+ },
1857
+ {
1858
+ "epoch": 0.6367851622874807,
1859
+ "grad_norm": 1.4827322959899902,
1860
+ "learning_rate": 1.3635239567233387e-05,
1861
+ "loss": 3.3267,
1862
+ "mean_token_accuracy": 0.4056967757642269,
1863
+ "num_tokens": 6494545.0,
1864
+ "step": 2060
1865
+ },
1866
+ {
1867
+ "epoch": 0.6398763523956723,
1868
+ "grad_norm": 2.8298745155334473,
1869
+ "learning_rate": 1.3604327666151469e-05,
1870
+ "loss": 3.2855,
1871
+ "mean_token_accuracy": 0.4048807807266712,
1872
+ "num_tokens": 6526119.0,
1873
+ "step": 2070
1874
+ },
1875
+ {
1876
+ "epoch": 0.642967542503864,
1877
+ "grad_norm": 2.294051170349121,
1878
+ "learning_rate": 1.3573415765069552e-05,
1879
+ "loss": 3.3445,
1880
+ "mean_token_accuracy": 0.3997214540839195,
1881
+ "num_tokens": 6559408.0,
1882
+ "step": 2080
1883
+ },
1884
+ {
1885
+ "epoch": 0.6460587326120556,
1886
+ "grad_norm": 6.12084436416626,
1887
+ "learning_rate": 1.3542503863987638e-05,
1888
+ "loss": 3.2882,
1889
+ "mean_token_accuracy": 0.40451241433620455,
1890
+ "num_tokens": 6588923.0,
1891
+ "step": 2090
1892
+ },
1893
+ {
1894
+ "epoch": 0.6491499227202473,
1895
+ "grad_norm": 1.5177021026611328,
1896
+ "learning_rate": 1.351159196290572e-05,
1897
+ "loss": 3.2833,
1898
+ "mean_token_accuracy": 0.39674848690629005,
1899
+ "num_tokens": 6623193.0,
1900
+ "step": 2100
1901
+ },
1902
+ {
1903
+ "epoch": 0.652241112828439,
1904
+ "grad_norm": 1.6658954620361328,
1905
+ "learning_rate": 1.3480680061823803e-05,
1906
+ "loss": 3.4093,
1907
+ "mean_token_accuracy": 0.39299999698996546,
1908
+ "num_tokens": 6657638.0,
1909
+ "step": 2110
1910
+ },
1911
+ {
1912
+ "epoch": 0.6553323029366306,
1913
+ "grad_norm": 5.038951396942139,
1914
+ "learning_rate": 1.3449768160741887e-05,
1915
+ "loss": 3.3359,
1916
+ "mean_token_accuracy": 0.4020788729190826,
1917
+ "num_tokens": 6686552.0,
1918
+ "step": 2120
1919
+ },
1920
+ {
1921
+ "epoch": 0.6584234930448223,
1922
+ "grad_norm": 1.931733250617981,
1923
+ "learning_rate": 1.3418856259659968e-05,
1924
+ "loss": 3.2761,
1925
+ "mean_token_accuracy": 0.4039643190801144,
1926
+ "num_tokens": 6719036.0,
1927
+ "step": 2130
1928
+ },
1929
+ {
1930
+ "epoch": 0.6615146831530139,
1931
+ "grad_norm": 1.7501503229141235,
1932
+ "learning_rate": 1.3387944358578054e-05,
1933
+ "loss": 3.2999,
1934
+ "mean_token_accuracy": 0.40199958309531214,
1935
+ "num_tokens": 6752457.0,
1936
+ "step": 2140
1937
+ },
1938
+ {
1939
+ "epoch": 0.6646058732612056,
1940
+ "grad_norm": 1.330138921737671,
1941
+ "learning_rate": 1.3357032457496137e-05,
1942
+ "loss": 3.2815,
1943
+ "mean_token_accuracy": 0.4096176542341709,
1944
+ "num_tokens": 6780937.0,
1945
+ "step": 2150
1946
+ },
1947
+ {
1948
+ "epoch": 0.6676970633693973,
1949
+ "grad_norm": 3.2849926948547363,
1950
+ "learning_rate": 1.3326120556414221e-05,
1951
+ "loss": 3.3212,
1952
+ "mean_token_accuracy": 0.39906698688864706,
1953
+ "num_tokens": 6813861.0,
1954
+ "step": 2160
1955
+ },
1956
+ {
1957
+ "epoch": 0.6707882534775889,
1958
+ "grad_norm": 1.3946915864944458,
1959
+ "learning_rate": 1.3295208655332303e-05,
1960
+ "loss": 3.3151,
1961
+ "mean_token_accuracy": 0.40061264783143996,
1962
+ "num_tokens": 6843256.0,
1963
+ "step": 2170
1964
+ },
1965
+ {
1966
+ "epoch": 0.6738794435857806,
1967
+ "grad_norm": 1.5147260427474976,
1968
+ "learning_rate": 1.3264296754250388e-05,
1969
+ "loss": 3.2324,
1970
+ "mean_token_accuracy": 0.4085581362247467,
1971
+ "num_tokens": 6881169.0,
1972
+ "step": 2180
1973
+ },
1974
+ {
1975
+ "epoch": 0.6769706336939721,
1976
+ "grad_norm": 1.5191727876663208,
1977
+ "learning_rate": 1.3233384853168472e-05,
1978
+ "loss": 3.2258,
1979
+ "mean_token_accuracy": 0.4069525547325611,
1980
+ "num_tokens": 6914374.0,
1981
+ "step": 2190
1982
+ },
1983
+ {
1984
+ "epoch": 0.6800618238021638,
1985
+ "grad_norm": 11.71318531036377,
1986
+ "learning_rate": 1.3202472952086554e-05,
1987
+ "loss": 3.2411,
1988
+ "mean_token_accuracy": 0.40940716192126275,
1989
+ "num_tokens": 6944810.0,
1990
+ "step": 2200
1991
+ },
1992
+ {
1993
+ "epoch": 0.6831530139103554,
1994
+ "grad_norm": 0.9575105309486389,
1995
+ "learning_rate": 1.3171561051004637e-05,
1996
+ "loss": 3.2721,
1997
+ "mean_token_accuracy": 0.40148399621248243,
1998
+ "num_tokens": 6975417.0,
1999
+ "step": 2210
2000
+ },
2001
+ {
2002
+ "epoch": 0.6862442040185471,
2003
+ "grad_norm": 1.6727248430252075,
2004
+ "learning_rate": 1.3140649149922722e-05,
2005
+ "loss": 3.3072,
2006
+ "mean_token_accuracy": 0.40622576996684073,
2007
+ "num_tokens": 7005387.0,
2008
+ "step": 2220
2009
+ },
2010
+ {
2011
+ "epoch": 0.6893353941267388,
2012
+ "grad_norm": 1.9732425212860107,
2013
+ "learning_rate": 1.3109737248840804e-05,
2014
+ "loss": 3.3486,
2015
+ "mean_token_accuracy": 0.4046866536140442,
2016
+ "num_tokens": 7036641.0,
2017
+ "step": 2230
2018
+ },
2019
+ {
2020
+ "epoch": 0.6924265842349304,
2021
+ "grad_norm": 3.0926458835601807,
2022
+ "learning_rate": 1.3078825347758888e-05,
2023
+ "loss": 3.2027,
2024
+ "mean_token_accuracy": 0.41145659387111666,
2025
+ "num_tokens": 7065613.0,
2026
+ "step": 2240
2027
+ },
2028
+ {
2029
+ "epoch": 0.6955177743431221,
2030
+ "grad_norm": 1.2291103601455688,
2031
+ "learning_rate": 1.3047913446676972e-05,
2032
+ "loss": 3.2273,
2033
+ "mean_token_accuracy": 0.41015600264072416,
2034
+ "num_tokens": 7098541.0,
2035
+ "step": 2250
2036
+ },
2037
+ {
2038
+ "epoch": 0.6986089644513137,
2039
+ "grad_norm": 1.393871784210205,
2040
+ "learning_rate": 1.3017001545595057e-05,
2041
+ "loss": 3.3534,
2042
+ "mean_token_accuracy": 0.398224713653326,
2043
+ "num_tokens": 7131991.0,
2044
+ "step": 2260
2045
+ },
2046
+ {
2047
+ "epoch": 0.7017001545595054,
2048
+ "grad_norm": 2.056251287460327,
2049
+ "learning_rate": 1.2986089644513139e-05,
2050
+ "loss": 3.2873,
2051
+ "mean_token_accuracy": 0.4067487485706806,
2052
+ "num_tokens": 7163563.0,
2053
+ "step": 2270
2054
+ },
2055
+ {
2056
+ "epoch": 0.7047913446676971,
2057
+ "grad_norm": 3.4474806785583496,
2058
+ "learning_rate": 1.2955177743431222e-05,
2059
+ "loss": 3.3688,
2060
+ "mean_token_accuracy": 0.39887151271104815,
2061
+ "num_tokens": 7196235.0,
2062
+ "step": 2280
2063
+ },
2064
+ {
2065
+ "epoch": 0.7078825347758887,
2066
+ "grad_norm": 4.399552345275879,
2067
+ "learning_rate": 1.2924265842349306e-05,
2068
+ "loss": 3.2107,
2069
+ "mean_token_accuracy": 0.4083859778940678,
2070
+ "num_tokens": 7229484.0,
2071
+ "step": 2290
2072
+ },
2073
+ {
2074
+ "epoch": 0.7109737248840804,
2075
+ "grad_norm": 1.0608441829681396,
2076
+ "learning_rate": 1.2893353941267388e-05,
2077
+ "loss": 3.3511,
2078
+ "mean_token_accuracy": 0.395867995172739,
2079
+ "num_tokens": 7263929.0,
2080
+ "step": 2300
2081
+ },
2082
+ {
2083
+ "epoch": 0.714064914992272,
2084
+ "grad_norm": 1.6351842880249023,
2085
+ "learning_rate": 1.2862442040185471e-05,
2086
+ "loss": 3.2312,
2087
+ "mean_token_accuracy": 0.40859498232603075,
2088
+ "num_tokens": 7297276.0,
2089
+ "step": 2310
2090
+ },
2091
+ {
2092
+ "epoch": 0.7171561051004637,
2093
+ "grad_norm": 1.5289595127105713,
2094
+ "learning_rate": 1.2831530139103557e-05,
2095
+ "loss": 3.325,
2096
+ "mean_token_accuracy": 0.40636713430285454,
2097
+ "num_tokens": 7329069.0,
2098
+ "step": 2320
2099
+ },
2100
+ {
2101
+ "epoch": 0.7202472952086554,
2102
+ "grad_norm": 1.5375980138778687,
2103
+ "learning_rate": 1.280061823802164e-05,
2104
+ "loss": 3.3243,
2105
+ "mean_token_accuracy": 0.40893488600850103,
2106
+ "num_tokens": 7362905.0,
2107
+ "step": 2330
2108
+ },
2109
+ {
2110
+ "epoch": 0.723338485316847,
2111
+ "grad_norm": 1.2137787342071533,
2112
+ "learning_rate": 1.2769706336939722e-05,
2113
+ "loss": 3.2944,
2114
+ "mean_token_accuracy": 0.4049150198698044,
2115
+ "num_tokens": 7392845.0,
2116
+ "step": 2340
2117
+ },
2118
+ {
2119
+ "epoch": 0.7264296754250387,
2120
+ "grad_norm": 3.618687152862549,
2121
+ "learning_rate": 1.2738794435857806e-05,
2122
+ "loss": 3.3114,
2123
+ "mean_token_accuracy": 0.398918454349041,
2124
+ "num_tokens": 7426273.0,
2125
+ "step": 2350
2126
+ },
2127
+ {
2128
+ "epoch": 0.7295208655332303,
2129
+ "grad_norm": 1.3081494569778442,
2130
+ "learning_rate": 1.2707882534775891e-05,
2131
+ "loss": 3.2284,
2132
+ "mean_token_accuracy": 0.4098616696894169,
2133
+ "num_tokens": 7457128.0,
2134
+ "step": 2360
2135
+ },
2136
+ {
2137
+ "epoch": 0.732612055641422,
2138
+ "grad_norm": 1.2867871522903442,
2139
+ "learning_rate": 1.2676970633693973e-05,
2140
+ "loss": 3.2734,
2141
+ "mean_token_accuracy": 0.40420192629098894,
2142
+ "num_tokens": 7490613.0,
2143
+ "step": 2370
2144
+ },
2145
+ {
2146
+ "epoch": 0.7357032457496137,
2147
+ "grad_norm": 6.16511869430542,
2148
+ "learning_rate": 1.2646058732612056e-05,
2149
+ "loss": 3.2275,
2150
+ "mean_token_accuracy": 0.4089609131217003,
2151
+ "num_tokens": 7524066.0,
2152
+ "step": 2380
2153
+ },
2154
+ {
2155
+ "epoch": 0.7387944358578052,
2156
+ "grad_norm": 3.3427209854125977,
2157
+ "learning_rate": 1.261514683153014e-05,
2158
+ "loss": 3.3428,
2159
+ "mean_token_accuracy": 0.3993862606585026,
2160
+ "num_tokens": 7554388.0,
2161
+ "step": 2390
2162
+ },
2163
+ {
2164
+ "epoch": 0.7418856259659969,
2165
+ "grad_norm": 1.759032964706421,
2166
+ "learning_rate": 1.2584234930448222e-05,
2167
+ "loss": 3.2427,
2168
+ "mean_token_accuracy": 0.411670895665884,
2169
+ "num_tokens": 7586805.0,
2170
+ "step": 2400
2171
+ },
2172
+ {
2173
+ "epoch": 0.7449768160741885,
2174
+ "grad_norm": 6.3605523109436035,
2175
+ "learning_rate": 1.2553323029366307e-05,
2176
+ "loss": 3.2989,
2177
+ "mean_token_accuracy": 0.4035753831267357,
2178
+ "num_tokens": 7621026.0,
2179
+ "step": 2410
2180
+ },
2181
+ {
2182
+ "epoch": 0.7480680061823802,
2183
+ "grad_norm": 1.3226512670516968,
2184
+ "learning_rate": 1.252241112828439e-05,
2185
+ "loss": 3.2676,
2186
+ "mean_token_accuracy": 0.4091234177350998,
2187
+ "num_tokens": 7653588.0,
2188
+ "step": 2420
2189
+ },
2190
+ {
2191
+ "epoch": 0.7511591962905718,
2192
+ "grad_norm": 1.2271887063980103,
2193
+ "learning_rate": 1.2491499227202474e-05,
2194
+ "loss": 3.1859,
2195
+ "mean_token_accuracy": 0.4189001992344856,
2196
+ "num_tokens": 7682672.0,
2197
+ "step": 2430
2198
+ },
2199
+ {
2200
+ "epoch": 0.7542503863987635,
2201
+ "grad_norm": 1.1789538860321045,
2202
+ "learning_rate": 1.2460587326120556e-05,
2203
+ "loss": 3.2707,
2204
+ "mean_token_accuracy": 0.41208377107977867,
2205
+ "num_tokens": 7711661.0,
2206
+ "step": 2440
2207
+ },
2208
+ {
2209
+ "epoch": 0.7573415765069552,
2210
+ "grad_norm": 1.4042397737503052,
2211
+ "learning_rate": 1.2429675425038642e-05,
2212
+ "loss": 3.3216,
2213
+ "mean_token_accuracy": 0.40250647664070127,
2214
+ "num_tokens": 7743853.0,
2215
+ "step": 2450
2216
+ },
2217
+ {
2218
+ "epoch": 0.7604327666151468,
2219
+ "grad_norm": 1.9280140399932861,
2220
+ "learning_rate": 1.2398763523956725e-05,
2221
+ "loss": 3.2948,
2222
+ "mean_token_accuracy": 0.4117195881903172,
2223
+ "num_tokens": 7776380.0,
2224
+ "step": 2460
2225
+ },
2226
+ {
2227
+ "epoch": 0.7635239567233385,
2228
+ "grad_norm": 1.244311809539795,
2229
+ "learning_rate": 1.2367851622874807e-05,
2230
+ "loss": 3.2397,
2231
+ "mean_token_accuracy": 0.41137626469135286,
2232
+ "num_tokens": 7808412.0,
2233
+ "step": 2470
2234
+ },
2235
+ {
2236
+ "epoch": 0.7666151468315301,
2237
+ "grad_norm": 1.2338584661483765,
2238
+ "learning_rate": 1.233693972179289e-05,
2239
+ "loss": 3.2887,
2240
+ "mean_token_accuracy": 0.41127968281507493,
2241
+ "num_tokens": 7837345.0,
2242
+ "step": 2480
2243
+ },
2244
+ {
2245
+ "epoch": 0.7697063369397218,
2246
+ "grad_norm": 1.0948349237442017,
2247
+ "learning_rate": 1.2306027820710976e-05,
2248
+ "loss": 3.2336,
2249
+ "mean_token_accuracy": 0.4080419853329659,
2250
+ "num_tokens": 7869621.0,
2251
+ "step": 2490
2252
+ },
2253
+ {
2254
+ "epoch": 0.7727975270479135,
2255
+ "grad_norm": 3.051591157913208,
2256
+ "learning_rate": 1.227511591962906e-05,
2257
+ "loss": 3.3946,
2258
+ "mean_token_accuracy": 0.39375910386443136,
2259
+ "num_tokens": 7900668.0,
2260
+ "step": 2500
2261
+ }
2262
+ ],
2263
+ "logging_steps": 10,
2264
+ "max_steps": 6470,
2265
+ "num_input_tokens_seen": 0,
2266
+ "num_train_epochs": 2,
2267
+ "save_steps": 500,
2268
+ "stateful_callbacks": {
2269
+ "TrainerControl": {
2270
+ "args": {
2271
+ "should_epoch_stop": false,
2272
+ "should_evaluate": false,
2273
+ "should_log": false,
2274
+ "should_save": true,
2275
+ "should_training_stop": false
2276
+ },
2277
+ "attributes": {}
2278
+ }
2279
+ },
2280
+ "total_flos": 4543540755056640.0,
2281
+ "train_batch_size": 4,
2282
+ "trial_name": null,
2283
+ "trial_params": null
2284
+ }
checkpoint-2500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7b555e5b9b497f109f303fcefbaa4f3565d3b5b6b83ddc274743aea488d2cd5
3
+ size 5624
checkpoint-3000/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": 256,
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-3000/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
+ }