Hugging Face's logo

Datasets:
PrompTartLAB
/
MUS-135-360

Dataset card Files
xet
Community
PrompTart commited on
Commit
ebddaea
·
verified ·
1 Parent(s): 5fabcba

Add files using upload-large-folder tool

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-1120/model.safetensors +3 -0
  2. checkpoint-1260/model.safetensors +3 -0
  3. checkpoint-1400/model.safetensors +3 -0
  4. checkpoint-1540/model.safetensors +3 -0
  5. checkpoint-1680/model.safetensors +3 -0
  6. checkpoint-1820/model.safetensors +3 -0
  7. checkpoint-1820/rng_state.pth +3 -0
  8. checkpoint-1820/scheduler.pt +3 -0
  9. checkpoint-1960/rng_state.pth +3 -0
  10. checkpoint-1960/scheduler.pt +3 -0
  11. checkpoint-1960/trainer_state.json +1828 -0
  12. checkpoint-1960/training_args.bin +3 -0
  13. checkpoint-2100/config.json +191 -0
  14. checkpoint-2100/merges.txt +0 -0
  15. checkpoint-2100/model.safetensors +3 -0
  16. checkpoint-2100/rng_state.pth +3 -0
  17. checkpoint-2100/scheduler.pt +3 -0
  18. checkpoint-2100/special_tokens_map.json +34 -0
  19. checkpoint-2100/tokenizer.json +0 -0
  20. checkpoint-2100/tokenizer_config.json +155 -0
  21. checkpoint-2100/trainer_state.json +1954 -0
  22. checkpoint-2100/training_args.bin +3 -0
  23. checkpoint-2100/vocab.json +0 -0
  24. checkpoint-2240/config.json +191 -0
  25. checkpoint-2240/merges.txt +0 -0
  26. checkpoint-2240/rng_state.pth +3 -0
  27. checkpoint-2240/scheduler.pt +3 -0
  28. checkpoint-2240/special_tokens_map.json +34 -0
  29. checkpoint-2240/tokenizer.json +0 -0
  30. checkpoint-2240/tokenizer_config.json +155 -0
  31. checkpoint-2240/trainer_state.json +2080 -0
  32. checkpoint-2240/training_args.bin +3 -0
  33. checkpoint-2240/vocab.json +0 -0
  34. checkpoint-2380/config.json +191 -0
  35. checkpoint-2380/merges.txt +0 -0
  36. checkpoint-2380/special_tokens_map.json +34 -0
  37. checkpoint-2380/tokenizer.json +0 -0
  38. checkpoint-2380/tokenizer_config.json +155 -0
  39. checkpoint-2380/trainer_state.json +2216 -0
  40. checkpoint-2380/vocab.json +0 -0
  41. checkpoint-2520/config.json +191 -0
  42. checkpoint-2520/merges.txt +0 -0
  43. checkpoint-2520/special_tokens_map.json +34 -0
  44. checkpoint-2520/tokenizer.json +0 -0
  45. checkpoint-2520/tokenizer_config.json +155 -0
  46. checkpoint-2520/trainer_state.json +2342 -0
  47. checkpoint-2520/vocab.json +0 -0
  48. checkpoint-2660/config.json +191 -0
  49. checkpoint-2660/merges.txt +0 -0
  50. checkpoint-2660/special_tokens_map.json +34 -0
checkpoint-1120/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd09d89a7497af4a5fc8bb533571dd93bf29ae1cb97bf90b5b75be3bc61019c5
3
+ size 2362909008
checkpoint-1260/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5b0418b0f5688517f41b8547665b81cd60f3bb0ff339c107ead43c8fbade7c8
3
+ size 2362909008
checkpoint-1400/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:657f0bd6678a900646e97a33f11a4a802f16c78cc5653bac4449cbf2ebe127ed
3
+ size 2362909008
checkpoint-1540/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc990e36044e5002cc056e69e0e96f521f0a8a29d35269969aa89c97e609e86b
3
+ size 2362909008
checkpoint-1680/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35cc4b9c1223984b67b524606bd707d0b3993a204397b07aa91067b78989ad65
3
+ size 2362909008
checkpoint-1820/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c105a49897023c19d3791318edd2c470d8404740c1758f8d8ac86821d5d0c11
3
+ size 2362909008
checkpoint-1820/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7d0e428eaaebbd39500df7b70c85e0c423ad7178fcc03acaa80860c51ef011a
3
+ size 14244
checkpoint-1820/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85be958d6c8157848e522465ec74a12db56807d004d0d56429782fb0d6f504fb
3
+ size 1064
checkpoint-1960/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7d0e428eaaebbd39500df7b70c85e0c423ad7178fcc03acaa80860c51ef011a
3
+ size 14244
checkpoint-1960/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5de1d86d52e5986e13e12474d1c3772242bab4952fe13d053f1d0587808cc337
3
+ size 1064
checkpoint-1960/trainer_state.json ADDED
@@ -0,0 +1,1828 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.4939812750780206,
6
+ "eval_steps": 500,
7
+ "global_step": 1960,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017833259028087384,
14
+ "grad_norm": 33.452999114990234,
15
+ "learning_rate": 3.2142857142857143e-06,
16
+ "loss": 8.9944,
17
+ "mean_token_accuracy": 0.32701094649964946,
18
+ "num_tokens": 2621440.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.03566651805617477,
23
+ "grad_norm": 2.596579074859619,
24
+ "learning_rate": 6.785714285714285e-06,
25
+ "loss": 4.9649,
26
+ "mean_token_accuracy": 0.5848119974136352,
27
+ "num_tokens": 5242880.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.05349977708426215,
32
+ "grad_norm": 1.8979898691177368,
33
+ "learning_rate": 1.0357142857142859e-05,
34
+ "loss": 4.1262,
35
+ "mean_token_accuracy": 0.5906830102205276,
36
+ "num_tokens": 7864320.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.07133303611234953,
41
+ "grad_norm": 1.1842058897018433,
42
+ "learning_rate": 1.392857142857143e-05,
43
+ "loss": 3.7327,
44
+ "mean_token_accuracy": 0.608327355235815,
45
+ "num_tokens": 10485760.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.08916629514043692,
50
+ "grad_norm": 0.922074556350708,
51
+ "learning_rate": 1.75e-05,
52
+ "loss": 3.4492,
53
+ "mean_token_accuracy": 0.6281662203371525,
54
+ "num_tokens": 13107200.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.1069995541685243,
59
+ "grad_norm": 0.7813541889190674,
60
+ "learning_rate": 2.107142857142857e-05,
61
+ "loss": 3.1416,
62
+ "mean_token_accuracy": 0.648241326212883,
63
+ "num_tokens": 15728640.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.12483281319661169,
68
+ "grad_norm": 1.1468125581741333,
69
+ "learning_rate": 2.4642857142857145e-05,
70
+ "loss": 2.8778,
71
+ "mean_token_accuracy": 0.6583510167896748,
72
+ "num_tokens": 18350080.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.14266607222469907,
77
+ "grad_norm": 1.0490431785583496,
78
+ "learning_rate": 2.8214285714285714e-05,
79
+ "loss": 2.7108,
80
+ "mean_token_accuracy": 0.663086223602295,
81
+ "num_tokens": 20971520.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.16049933125278645,
86
+ "grad_norm": 0.7813755869865417,
87
+ "learning_rate": 3.1785714285714284e-05,
88
+ "loss": 2.6039,
89
+ "mean_token_accuracy": 0.6650082431733608,
90
+ "num_tokens": 23592960.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.17833259028087384,
95
+ "grad_norm": 0.6500654220581055,
96
+ "learning_rate": 3.5357142857142864e-05,
97
+ "loss": 2.4513,
98
+ "mean_token_accuracy": 0.674950385093689,
99
+ "num_tokens": 26214400.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.19616584930896122,
104
+ "grad_norm": 0.81620192527771,
105
+ "learning_rate": 3.892857142857143e-05,
106
+ "loss": 2.3602,
107
+ "mean_token_accuracy": 0.6777540303766727,
108
+ "num_tokens": 28835840.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.2139991083370486,
113
+ "grad_norm": 0.6131238341331482,
114
+ "learning_rate": 4.25e-05,
115
+ "loss": 2.3545,
116
+ "mean_token_accuracy": 0.6758881151676178,
117
+ "num_tokens": 31457280.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.231832367365136,
122
+ "grad_norm": 0.9951689839363098,
123
+ "learning_rate": 4.607142857142857e-05,
124
+ "loss": 2.2624,
125
+ "mean_token_accuracy": 0.6806462205946445,
126
+ "num_tokens": 34078720.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.24966562639322337,
131
+ "grad_norm": 0.7796569466590881,
132
+ "learning_rate": 4.964285714285715e-05,
133
+ "loss": 2.1505,
134
+ "mean_token_accuracy": 0.6918394304811954,
135
+ "num_tokens": 36700160.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.26749888542131073,
140
+ "grad_norm": 0.8614610433578491,
141
+ "learning_rate": 4.9998587698517684e-05,
142
+ "loss": 2.1591,
143
+ "mean_token_accuracy": 0.68697941750288,
144
+ "num_tokens": 39321600.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.28533214444939814,
149
+ "grad_norm": 0.6178464889526367,
150
+ "learning_rate": 4.9993705873562665e-05,
151
+ "loss": 2.1048,
152
+ "mean_token_accuracy": 0.6911555036902428,
153
+ "num_tokens": 41943040.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.3031654034774855,
158
+ "grad_norm": 0.7255717515945435,
159
+ "learning_rate": 4.998533777009954e-05,
160
+ "loss": 2.0847,
161
+ "mean_token_accuracy": 0.6920825429260731,
162
+ "num_tokens": 44564480.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.3209986625055729,
167
+ "grad_norm": 0.7800336480140686,
168
+ "learning_rate": 4.9973484555363726e-05,
169
+ "loss": 2.049,
170
+ "mean_token_accuracy": 0.6982546880841255,
171
+ "num_tokens": 47185920.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.33883192153366026,
176
+ "grad_norm": 0.886745274066925,
177
+ "learning_rate": 4.995814788271582e-05,
178
+ "loss": 2.0301,
179
+ "mean_token_accuracy": 0.697366189956665,
180
+ "num_tokens": 49807360.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.3566651805617477,
185
+ "grad_norm": 0.6190801858901978,
186
+ "learning_rate": 4.993932989141088e-05,
187
+ "loss": 1.9408,
188
+ "mean_token_accuracy": 0.7095360569655895,
189
+ "num_tokens": 52428800.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.37449843958983503,
194
+ "grad_norm": 0.6333149671554565,
195
+ "learning_rate": 4.991703320630011e-05,
196
+ "loss": 1.9076,
197
+ "mean_token_accuracy": 0.7127362459897995,
198
+ "num_tokens": 55050240.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.39233169861792244,
203
+ "grad_norm": 0.9046627879142761,
204
+ "learning_rate": 4.989126093746468e-05,
205
+ "loss": 1.8913,
206
+ "mean_token_accuracy": 0.7148448958992958,
207
+ "num_tokens": 57671680.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.4101649576460098,
212
+ "grad_norm": 0.7778757214546204,
213
+ "learning_rate": 4.986201667978193e-05,
214
+ "loss": 1.9016,
215
+ "mean_token_accuracy": 0.7118340134620667,
216
+ "num_tokens": 60293120.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.4279982166740972,
221
+ "grad_norm": 0.6346383690834045,
222
+ "learning_rate": 4.9829304512423944e-05,
223
+ "loss": 1.898,
224
+ "mean_token_accuracy": 0.7123855009675026,
225
+ "num_tokens": 62914560.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.44583147570218457,
230
+ "grad_norm": 0.7308981418609619,
231
+ "learning_rate": 4.979312899828853e-05,
232
+ "loss": 1.8435,
233
+ "mean_token_accuracy": 0.7184088006615639,
234
+ "num_tokens": 65536000.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.463664734730272,
239
+ "grad_norm": 0.6174861788749695,
240
+ "learning_rate": 4.9753495183362796e-05,
241
+ "loss": 1.8242,
242
+ "mean_token_accuracy": 0.7238672457635402,
243
+ "num_tokens": 68157440.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.48149799375835933,
248
+ "grad_norm": 0.6507575511932373,
249
+ "learning_rate": 4.971040859601927e-05,
250
+ "loss": 1.791,
251
+ "mean_token_accuracy": 0.7251060992479325,
252
+ "num_tokens": 70778880.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.49933125278644674,
257
+ "grad_norm": 0.8092382550239563,
258
+ "learning_rate": 4.966387524624482e-05,
259
+ "loss": 1.7879,
260
+ "mean_token_accuracy": 0.7255102708935738,
261
+ "num_tokens": 73400320.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.5171645118145342,
266
+ "grad_norm": 0.5636172294616699,
267
+ "learning_rate": 4.9613901624802245e-05,
268
+ "loss": 1.7351,
269
+ "mean_token_accuracy": 0.7330663032829762,
270
+ "num_tokens": 76021760.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.5349977708426215,
275
+ "grad_norm": 0.655262291431427,
276
+ "learning_rate": 4.956049470232503e-05,
277
+ "loss": 1.7464,
278
+ "mean_token_accuracy": 0.7302443347871304,
279
+ "num_tokens": 78643200.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.5528310298707089,
284
+ "grad_norm": 0.9413332343101501,
285
+ "learning_rate": 4.9503661928344944e-05,
286
+ "loss": 1.7277,
287
+ "mean_token_accuracy": 0.730958030372858,
288
+ "num_tokens": 81264640.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.5706642888987963,
293
+ "grad_norm": 0.7439999580383301,
294
+ "learning_rate": 4.9443411230252956e-05,
295
+ "loss": 1.6947,
296
+ "mean_token_accuracy": 0.7346593342721462,
297
+ "num_tokens": 83886080.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.5884975479268837,
302
+ "grad_norm": 0.7665299773216248,
303
+ "learning_rate": 4.93797510121935e-05,
304
+ "loss": 1.7136,
305
+ "mean_token_accuracy": 0.7324949607253075,
306
+ "num_tokens": 86507520.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.606330806954971,
311
+ "grad_norm": 0.7366372346878052,
312
+ "learning_rate": 4.931269015389217e-05,
313
+ "loss": 1.676,
314
+ "mean_token_accuracy": 0.7371297933161258,
315
+ "num_tokens": 89128960.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.6241640659830584,
320
+ "grad_norm": 0.6571912169456482,
321
+ "learning_rate": 4.9242238009417175e-05,
322
+ "loss": 1.6252,
323
+ "mean_token_accuracy": 0.7434912383556366,
324
+ "num_tokens": 91750400.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.6419973250111458,
329
+ "grad_norm": 0.7496657967567444,
330
+ "learning_rate": 4.9168404405874505e-05,
331
+ "loss": 1.657,
332
+ "mean_token_accuracy": 0.737925547361374,
333
+ "num_tokens": 94371840.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.6598305840392331,
338
+ "grad_norm": 0.5275835394859314,
339
+ "learning_rate": 4.9091199642037244e-05,
340
+ "loss": 1.6723,
341
+ "mean_token_accuracy": 0.7358783438801766,
342
+ "num_tokens": 96993280.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.6776638430673205,
347
+ "grad_norm": 0.6143308281898499,
348
+ "learning_rate": 4.901063448690901e-05,
349
+ "loss": 1.6292,
350
+ "mean_token_accuracy": 0.7412570238113403,
351
+ "num_tokens": 99614720.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.6954971020954079,
356
+ "grad_norm": 0.860007643699646,
357
+ "learning_rate": 4.892672017822183e-05,
358
+ "loss": 1.6281,
359
+ "mean_token_accuracy": 0.7422710858285427,
360
+ "num_tokens": 102236160.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.7133303611234953,
365
+ "grad_norm": 0.5295814275741577,
366
+ "learning_rate": 4.8839468420868606e-05,
367
+ "loss": 1.6112,
368
+ "mean_token_accuracy": 0.7446457460522652,
369
+ "num_tokens": 104857600.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.7311636201515826,
374
+ "grad_norm": 0.7366165518760681,
375
+ "learning_rate": 4.8748891385270495e-05,
376
+ "loss": 1.6197,
377
+ "mean_token_accuracy": 0.7417318046092987,
378
+ "num_tokens": 107479040.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.7489968791796701,
383
+ "grad_norm": 0.7033573985099792,
384
+ "learning_rate": 4.865500170567926e-05,
385
+ "loss": 1.5763,
386
+ "mean_token_accuracy": 0.7474127545952797,
387
+ "num_tokens": 110100480.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.7668301382077575,
392
+ "grad_norm": 0.6527330875396729,
393
+ "learning_rate": 4.855781247841498e-05,
394
+ "loss": 1.5714,
395
+ "mean_token_accuracy": 0.7477886863052845,
396
+ "num_tokens": 112721920.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.7846633972358449,
401
+ "grad_norm": 0.7344551682472229,
402
+ "learning_rate": 4.845733726003928e-05,
403
+ "loss": 1.5682,
404
+ "mean_token_accuracy": 0.7475516766309738,
405
+ "num_tokens": 115343360.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.8024966562639322,
410
+ "grad_norm": 0.6840972304344177,
411
+ "learning_rate": 4.835359006546443e-05,
412
+ "loss": 1.5955,
413
+ "mean_token_accuracy": 0.7427977688610554,
414
+ "num_tokens": 117964800.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.8203299152920196,
419
+ "grad_norm": 0.635522723197937,
420
+ "learning_rate": 4.824658536599836e-05,
421
+ "loss": 1.5459,
422
+ "mean_token_accuracy": 0.7526196874678135,
423
+ "num_tokens": 120586240.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.838163174320107,
428
+ "grad_norm": 0.6874827146530151,
429
+ "learning_rate": 4.8136338087326216e-05,
430
+ "loss": 1.6108,
431
+ "mean_token_accuracy": 0.7413325920701027,
432
+ "num_tokens": 123207680.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.8559964333481944,
437
+ "grad_norm": 0.6571274995803833,
438
+ "learning_rate": 4.8022863607428356e-05,
439
+ "loss": 1.539,
440
+ "mean_token_accuracy": 0.7521994858980179,
441
+ "num_tokens": 125829120.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.8738296923762817,
446
+ "grad_norm": 0.8841336369514465,
447
+ "learning_rate": 4.790617775443537e-05,
448
+ "loss": 1.5486,
449
+ "mean_token_accuracy": 0.7492702730000019,
450
+ "num_tokens": 128450560.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.8916629514043691,
455
+ "grad_norm": 0.690990149974823,
456
+ "learning_rate": 4.7786296804420274e-05,
457
+ "loss": 1.5646,
458
+ "mean_token_accuracy": 0.7455468386411667,
459
+ "num_tokens": 131072000.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.9094962104324565,
464
+ "grad_norm": 0.7215222716331482,
465
+ "learning_rate": 4.766323747912818e-05,
466
+ "loss": 1.5085,
467
+ "mean_token_accuracy": 0.754824897646904,
468
+ "num_tokens": 133693440.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.927329469460544,
473
+ "grad_norm": 0.6024391651153564,
474
+ "learning_rate": 4.753701694364392e-05,
475
+ "loss": 1.4856,
476
+ "mean_token_accuracy": 0.7570274345576763,
477
+ "num_tokens": 136314880.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.9451627284886313,
482
+ "grad_norm": 0.6131455302238464,
483
+ "learning_rate": 4.740765280399767e-05,
484
+ "loss": 1.5433,
485
+ "mean_token_accuracy": 0.7495988786220551,
486
+ "num_tokens": 138936320.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.9629959875167187,
491
+ "grad_norm": 0.7875815629959106,
492
+ "learning_rate": 4.72751631047092e-05,
493
+ "loss": 1.4917,
494
+ "mean_token_accuracy": 0.7559954360127449,
495
+ "num_tokens": 141557760.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.9808292465448061,
500
+ "grad_norm": 0.7112794518470764,
501
+ "learning_rate": 4.71395663262709e-05,
502
+ "loss": 1.4784,
503
+ "mean_token_accuracy": 0.7572461232542992,
504
+ "num_tokens": 144179200.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.9986625055728935,
509
+ "grad_norm": 0.6807007789611816,
510
+ "learning_rate": 4.7000881382570014e-05,
511
+ "loss": 1.5159,
512
+ "mean_token_accuracy": 0.7523536741733551,
513
+ "num_tokens": 146800640.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 1.0,
518
+ "eval_loss": 1.494895339012146,
519
+ "eval_mean_token_accuracy": 0.7405715703964233,
520
+ "eval_num_tokens": 146980864.0,
521
+ "eval_runtime": 2.6056,
522
+ "eval_samples_per_second": 27.633,
523
+ "eval_steps_per_second": 1.919,
524
+ "step": 561
525
+ },
526
+ {
527
+ "epoch": 1.0160499331252786,
528
+ "grad_norm": 0.5480894446372986,
529
+ "learning_rate": 4.685912761825038e-05,
530
+ "loss": 1.4428,
531
+ "mean_token_accuracy": 0.7618689170250525,
532
+ "num_tokens": 149340160.0,
533
+ "step": 570
534
+ },
535
+ {
536
+ "epoch": 1.033883192153366,
537
+ "grad_norm": 0.6192435622215271,
538
+ "learning_rate": 4.671432480601416e-05,
539
+ "loss": 1.4711,
540
+ "mean_token_accuracy": 0.7556988887488842,
541
+ "num_tokens": 151961600.0,
542
+ "step": 580
543
+ },
544
+ {
545
+ "epoch": 1.0517164511814534,
546
+ "grad_norm": 0.5988848209381104,
547
+ "learning_rate": 4.656649314386378e-05,
548
+ "loss": 1.433,
549
+ "mean_token_accuracy": 0.7611474089324475,
550
+ "num_tokens": 154583040.0,
551
+ "step": 590
552
+ },
553
+ {
554
+ "epoch": 1.0695497102095408,
555
+ "grad_norm": 0.7051531672477722,
556
+ "learning_rate": 4.641565325228464e-05,
557
+ "loss": 1.415,
558
+ "mean_token_accuracy": 0.7643376767635346,
559
+ "num_tokens": 157204480.0,
560
+ "step": 600
561
+ },
562
+ {
563
+ "epoch": 1.0873829692376282,
564
+ "grad_norm": 0.6170579791069031,
565
+ "learning_rate": 4.6261826171368774e-05,
566
+ "loss": 1.4674,
567
+ "mean_token_accuracy": 0.7571297205984593,
568
+ "num_tokens": 159825920.0,
569
+ "step": 610
570
+ },
571
+ {
572
+ "epoch": 1.1052162282657156,
573
+ "grad_norm": 0.5691844820976257,
574
+ "learning_rate": 4.6105033357880104e-05,
575
+ "loss": 1.407,
576
+ "mean_token_accuracy": 0.7646078862249851,
577
+ "num_tokens": 162447360.0,
578
+ "step": 620
579
+ },
580
+ {
581
+ "epoch": 1.123049487293803,
582
+ "grad_norm": 0.5824171900749207,
583
+ "learning_rate": 4.594529668226146e-05,
584
+ "loss": 1.4185,
585
+ "mean_token_accuracy": 0.7614389970898628,
586
+ "num_tokens": 165068800.0,
587
+ "step": 630
588
+ },
589
+ {
590
+ "epoch": 1.1408827463218902,
591
+ "grad_norm": 0.5765758156776428,
592
+ "learning_rate": 4.578263842558402e-05,
593
+ "loss": 1.42,
594
+ "mean_token_accuracy": 0.7633033894002438,
595
+ "num_tokens": 167690240.0,
596
+ "step": 640
597
+ },
598
+ {
599
+ "epoch": 1.1587160053499777,
600
+ "grad_norm": 0.5261256694793701,
601
+ "learning_rate": 4.5617081276439355e-05,
602
+ "loss": 1.4384,
603
+ "mean_token_accuracy": 0.7604428693652153,
604
+ "num_tokens": 170311680.0,
605
+ "step": 650
606
+ },
607
+ {
608
+ "epoch": 1.176549264378065,
609
+ "grad_norm": 0.913208544254303,
610
+ "learning_rate": 4.544864832777469e-05,
611
+ "loss": 1.4775,
612
+ "mean_token_accuracy": 0.754877183586359,
613
+ "num_tokens": 172933120.0,
614
+ "step": 660
615
+ },
616
+ {
617
+ "epoch": 1.1943825234061525,
618
+ "grad_norm": 0.5799976587295532,
619
+ "learning_rate": 4.527736307367179e-05,
620
+ "loss": 1.4387,
621
+ "mean_token_accuracy": 0.7589173935353756,
622
+ "num_tokens": 175554560.0,
623
+ "step": 670
624
+ },
625
+ {
626
+ "epoch": 1.2122157824342399,
627
+ "grad_norm": 0.635047435760498,
628
+ "learning_rate": 4.510324940606979e-05,
629
+ "loss": 1.4439,
630
+ "mean_token_accuracy": 0.7597284145653248,
631
+ "num_tokens": 178176000.0,
632
+ "step": 680
633
+ },
634
+ {
635
+ "epoch": 1.2300490414623273,
636
+ "grad_norm": 0.5884873270988464,
637
+ "learning_rate": 4.4926331611432687e-05,
638
+ "loss": 1.3835,
639
+ "mean_token_accuracy": 0.7659787967801094,
640
+ "num_tokens": 180797440.0,
641
+ "step": 690
642
+ },
643
+ {
644
+ "epoch": 1.2478823004904147,
645
+ "grad_norm": 0.7232898473739624,
646
+ "learning_rate": 4.4746634367361634e-05,
647
+ "loss": 1.3629,
648
+ "mean_token_accuracy": 0.7691854700446129,
649
+ "num_tokens": 183418880.0,
650
+ "step": 700
651
+ },
652
+ {
653
+ "epoch": 1.265715559518502,
654
+ "grad_norm": 0.6569238305091858,
655
+ "learning_rate": 4.4564182739152805e-05,
656
+ "loss": 1.397,
657
+ "mean_token_accuracy": 0.7651830442249775,
658
+ "num_tokens": 186040320.0,
659
+ "step": 710
660
+ },
661
+ {
662
+ "epoch": 1.2835488185465893,
663
+ "grad_norm": 0.8630965352058411,
664
+ "learning_rate": 4.437900217630109e-05,
665
+ "loss": 1.4256,
666
+ "mean_token_accuracy": 0.7605882868170738,
667
+ "num_tokens": 188661760.0,
668
+ "step": 720
669
+ },
670
+ {
671
+ "epoch": 1.3013820775746767,
672
+ "grad_norm": 0.7600955367088318,
673
+ "learning_rate": 4.419111850895028e-05,
674
+ "loss": 1.3975,
675
+ "mean_token_accuracy": 0.7634797111153603,
676
+ "num_tokens": 191283200.0,
677
+ "step": 730
678
+ },
679
+ {
680
+ "epoch": 1.3192153366027641,
681
+ "grad_norm": 0.6687177419662476,
682
+ "learning_rate": 4.4000557944290086e-05,
683
+ "loss": 1.3616,
684
+ "mean_token_accuracy": 0.7703544825315476,
685
+ "num_tokens": 193904640.0,
686
+ "step": 740
687
+ },
688
+ {
689
+ "epoch": 1.3370485956308515,
690
+ "grad_norm": 0.719222903251648,
691
+ "learning_rate": 4.3807347062900624e-05,
692
+ "loss": 1.3788,
693
+ "mean_token_accuracy": 0.7648544363677502,
694
+ "num_tokens": 196526080.0,
695
+ "step": 750
696
+ },
697
+ {
698
+ "epoch": 1.354881854658939,
699
+ "grad_norm": 0.5986538529396057,
700
+ "learning_rate": 4.361151281504474e-05,
701
+ "loss": 1.364,
702
+ "mean_token_accuracy": 0.770479665696621,
703
+ "num_tokens": 199147520.0,
704
+ "step": 760
705
+ },
706
+ {
707
+ "epoch": 1.3727151136870264,
708
+ "grad_norm": 0.6472992897033691,
709
+ "learning_rate": 4.3413082516908893e-05,
710
+ "loss": 1.4045,
711
+ "mean_token_accuracy": 0.7644098080694676,
712
+ "num_tokens": 201768960.0,
713
+ "step": 770
714
+ },
715
+ {
716
+ "epoch": 1.3905483727151138,
717
+ "grad_norm": 0.44522207975387573,
718
+ "learning_rate": 4.321208384679285e-05,
719
+ "loss": 1.3314,
720
+ "mean_token_accuracy": 0.773588253557682,
721
+ "num_tokens": 204390400.0,
722
+ "step": 780
723
+ },
724
+ {
725
+ "epoch": 1.4083816317432012,
726
+ "grad_norm": 0.5019171833992004,
727
+ "learning_rate": 4.3008544841248986e-05,
728
+ "loss": 1.3591,
729
+ "mean_token_accuracy": 0.7695087313652038,
730
+ "num_tokens": 207011840.0,
731
+ "step": 790
732
+ },
733
+ {
734
+ "epoch": 1.4262148907712884,
735
+ "grad_norm": 0.7287567257881165,
736
+ "learning_rate": 4.280249389117157e-05,
737
+ "loss": 1.3773,
738
+ "mean_token_accuracy": 0.7675351880490779,
739
+ "num_tokens": 209633280.0,
740
+ "step": 800
741
+ },
742
+ {
743
+ "epoch": 1.4440481497993758,
744
+ "grad_norm": 1.0294618606567383,
745
+ "learning_rate": 4.2593959737836604e-05,
746
+ "loss": 1.3719,
747
+ "mean_token_accuracy": 0.7668005004525185,
748
+ "num_tokens": 212254720.0,
749
+ "step": 810
750
+ },
751
+ {
752
+ "epoch": 1.4618814088274632,
753
+ "grad_norm": 0.6181490421295166,
754
+ "learning_rate": 4.2382971468892806e-05,
755
+ "loss": 1.3168,
756
+ "mean_token_accuracy": 0.7758259050548076,
757
+ "num_tokens": 214876160.0,
758
+ "step": 820
759
+ },
760
+ {
761
+ "epoch": 1.4797146678555506,
762
+ "grad_norm": 0.5945810079574585,
763
+ "learning_rate": 4.216955851430432e-05,
764
+ "loss": 1.3536,
765
+ "mean_token_accuracy": 0.7710937492549419,
766
+ "num_tokens": 217497600.0,
767
+ "step": 830
768
+ },
769
+ {
770
+ "epoch": 1.497547926883638,
771
+ "grad_norm": 0.6803934574127197,
772
+ "learning_rate": 4.195375064224562e-05,
773
+ "loss": 1.324,
774
+ "mean_token_accuracy": 0.7753530308604241,
775
+ "num_tokens": 220119040.0,
776
+ "step": 840
777
+ },
778
+ {
779
+ "epoch": 1.5153811859117252,
780
+ "grad_norm": 0.5748502612113953,
781
+ "learning_rate": 4.1735577954949225e-05,
782
+ "loss": 1.3057,
783
+ "mean_token_accuracy": 0.7755480580031872,
784
+ "num_tokens": 222740480.0,
785
+ "step": 850
786
+ },
787
+ {
788
+ "epoch": 1.5332144449398126,
789
+ "grad_norm": 0.5675086379051208,
790
+ "learning_rate": 4.151507088450692e-05,
791
+ "loss": 1.302,
792
+ "mean_token_accuracy": 0.7765964694321156,
793
+ "num_tokens": 225361920.0,
794
+ "step": 860
795
+ },
796
+ {
797
+ "epoch": 1.5510477039679,
798
+ "grad_norm": 0.6266924142837524,
799
+ "learning_rate": 4.129226018862484e-05,
800
+ "loss": 1.353,
801
+ "mean_token_accuracy": 0.770804837346077,
802
+ "num_tokens": 227983360.0,
803
+ "step": 870
804
+ },
805
+ {
806
+ "epoch": 1.5688809629959874,
807
+ "grad_norm": 0.551527202129364,
808
+ "learning_rate": 4.106717694633321e-05,
809
+ "loss": 1.3477,
810
+ "mean_token_accuracy": 0.77206010222435,
811
+ "num_tokens": 230604800.0,
812
+ "step": 880
813
+ },
814
+ {
815
+ "epoch": 1.5867142220240749,
816
+ "grad_norm": 0.6628434658050537,
817
+ "learning_rate": 4.0839852553651265e-05,
818
+ "loss": 1.3244,
819
+ "mean_token_accuracy": 0.773486353456974,
820
+ "num_tokens": 233226240.0,
821
+ "step": 890
822
+ },
823
+ {
824
+ "epoch": 1.6045474810521623,
825
+ "grad_norm": 0.5956223011016846,
826
+ "learning_rate": 4.0610318719207906e-05,
827
+ "loss": 1.3298,
828
+ "mean_token_accuracy": 0.7733359761536122,
829
+ "num_tokens": 235847680.0,
830
+ "step": 900
831
+ },
832
+ {
833
+ "epoch": 1.6223807400802497,
834
+ "grad_norm": 0.6573621034622192,
835
+ "learning_rate": 4.037860745981881e-05,
836
+ "loss": 1.283,
837
+ "mean_token_accuracy": 0.7787142746150494,
838
+ "num_tokens": 238469120.0,
839
+ "step": 910
840
+ },
841
+ {
842
+ "epoch": 1.640213999108337,
843
+ "grad_norm": 0.607742965221405,
844
+ "learning_rate": 4.01447510960205e-05,
845
+ "loss": 1.2955,
846
+ "mean_token_accuracy": 0.7780937045812607,
847
+ "num_tokens": 241090560.0,
848
+ "step": 920
849
+ },
850
+ {
851
+ "epoch": 1.6580472581364245,
852
+ "grad_norm": 0.6278334856033325,
853
+ "learning_rate": 3.99087822475621e-05,
854
+ "loss": 1.3215,
855
+ "mean_token_accuracy": 0.7741382211446762,
856
+ "num_tokens": 243712000.0,
857
+ "step": 930
858
+ },
859
+ {
860
+ "epoch": 1.675880517164512,
861
+ "grad_norm": 0.5402811169624329,
862
+ "learning_rate": 3.96707338288553e-05,
863
+ "loss": 1.3113,
864
+ "mean_token_accuracy": 0.7751454092562199,
865
+ "num_tokens": 246333440.0,
866
+ "step": 940
867
+ },
868
+ {
869
+ "epoch": 1.6937137761925993,
870
+ "grad_norm": 0.6165571212768555,
871
+ "learning_rate": 3.9430639044383274e-05,
872
+ "loss": 1.3093,
873
+ "mean_token_accuracy": 0.7768628627061844,
874
+ "num_tokens": 248954880.0,
875
+ "step": 950
876
+ },
877
+ {
878
+ "epoch": 1.7115470352206867,
879
+ "grad_norm": 0.6086046695709229,
880
+ "learning_rate": 3.9188531384069096e-05,
881
+ "loss": 1.3125,
882
+ "mean_token_accuracy": 0.7762621335685254,
883
+ "num_tokens": 251576320.0,
884
+ "step": 960
885
+ },
886
+ {
887
+ "epoch": 1.729380294248774,
888
+ "grad_norm": 0.682686984539032,
889
+ "learning_rate": 3.8944444618604356e-05,
890
+ "loss": 1.2821,
891
+ "mean_token_accuracy": 0.7795726217329502,
892
+ "num_tokens": 254197760.0,
893
+ "step": 970
894
+ },
895
+ {
896
+ "epoch": 1.7472135532768613,
897
+ "grad_norm": 0.5615099668502808,
898
+ "learning_rate": 3.869841279473862e-05,
899
+ "loss": 1.3095,
900
+ "mean_token_accuracy": 0.7766918800771236,
901
+ "num_tokens": 256819200.0,
902
+ "step": 980
903
+ },
904
+ {
905
+ "epoch": 1.7650468123049488,
906
+ "grad_norm": 0.5341463088989258,
907
+ "learning_rate": 3.845047023053037e-05,
908
+ "loss": 1.2864,
909
+ "mean_token_accuracy": 0.779373399913311,
910
+ "num_tokens": 259440640.0,
911
+ "step": 990
912
+ },
913
+ {
914
+ "epoch": 1.7828800713330362,
915
+ "grad_norm": 0.5641157627105713,
916
+ "learning_rate": 3.82006515105601e-05,
917
+ "loss": 1.2857,
918
+ "mean_token_accuracy": 0.7798695519566536,
919
+ "num_tokens": 262062080.0,
920
+ "step": 1000
921
+ },
922
+ {
923
+ "epoch": 1.8007133303611234,
924
+ "grad_norm": 0.5196830630302429,
925
+ "learning_rate": 3.794899148110628e-05,
926
+ "loss": 1.2949,
927
+ "mean_token_accuracy": 0.7788909852504731,
928
+ "num_tokens": 264683520.0,
929
+ "step": 1010
930
+ },
931
+ {
932
+ "epoch": 1.8185465893892108,
933
+ "grad_norm": 0.4990008771419525,
934
+ "learning_rate": 3.7695525245284716e-05,
935
+ "loss": 1.254,
936
+ "mean_token_accuracy": 0.7839147254824639,
937
+ "num_tokens": 267304960.0,
938
+ "step": 1020
939
+ },
940
+ {
941
+ "epoch": 1.8363798484172982,
942
+ "grad_norm": 0.5643882751464844,
943
+ "learning_rate": 3.7440288158152187e-05,
944
+ "loss": 1.2446,
945
+ "mean_token_accuracy": 0.785726061463356,
946
+ "num_tokens": 269926400.0,
947
+ "step": 1030
948
+ },
949
+ {
950
+ "epoch": 1.8542131074453856,
951
+ "grad_norm": 0.8107313513755798,
952
+ "learning_rate": 3.7183315821774914e-05,
953
+ "loss": 1.2974,
954
+ "mean_token_accuracy": 0.7775464087724686,
955
+ "num_tokens": 272547840.0,
956
+ "step": 1040
957
+ },
958
+ {
959
+ "epoch": 1.872046366473473,
960
+ "grad_norm": 0.5176706314086914,
961
+ "learning_rate": 3.692464408026253e-05,
962
+ "loss": 1.2916,
963
+ "mean_token_accuracy": 0.7772151328623295,
964
+ "num_tokens": 275169280.0,
965
+ "step": 1050
966
+ },
967
+ {
968
+ "epoch": 1.8898796255015604,
969
+ "grad_norm": 0.5898165106773376,
970
+ "learning_rate": 3.666430901476833e-05,
971
+ "loss": 1.2716,
972
+ "mean_token_accuracy": 0.7806546173989772,
973
+ "num_tokens": 277790720.0,
974
+ "step": 1060
975
+ },
976
+ {
977
+ "epoch": 1.9077128845296478,
978
+ "grad_norm": 0.5368467569351196,
979
+ "learning_rate": 3.640234693845645e-05,
980
+ "loss": 1.2491,
981
+ "mean_token_accuracy": 0.7844746112823486,
982
+ "num_tokens": 280412160.0,
983
+ "step": 1070
984
+ },
985
+ {
986
+ "epoch": 1.9255461435577352,
987
+ "grad_norm": 0.5310569405555725,
988
+ "learning_rate": 3.613879439143667e-05,
989
+ "loss": 1.2825,
990
+ "mean_token_accuracy": 0.7795199528336525,
991
+ "num_tokens": 283033600.0,
992
+ "step": 1080
993
+ },
994
+ {
995
+ "epoch": 1.9433794025858226,
996
+ "grad_norm": 0.6114910840988159,
997
+ "learning_rate": 3.58736881356676e-05,
998
+ "loss": 1.2595,
999
+ "mean_token_accuracy": 0.7831220239400863,
1000
+ "num_tokens": 285655040.0,
1001
+ "step": 1090
1002
+ },
1003
+ {
1004
+ "epoch": 1.96121266161391,
1005
+ "grad_norm": 0.5786811113357544,
1006
+ "learning_rate": 3.5607065149828843e-05,
1007
+ "loss": 1.2694,
1008
+ "mean_token_accuracy": 0.7822808526456356,
1009
+ "num_tokens": 288276480.0,
1010
+ "step": 1100
1011
+ },
1012
+ {
1013
+ "epoch": 1.9790459206419975,
1014
+ "grad_norm": 0.6175279021263123,
1015
+ "learning_rate": 3.533896262416302e-05,
1016
+ "loss": 1.2712,
1017
+ "mean_token_accuracy": 0.7803981445729733,
1018
+ "num_tokens": 290897920.0,
1019
+ "step": 1110
1020
+ },
1021
+ {
1022
+ "epoch": 1.9968791796700847,
1023
+ "grad_norm": 0.568803608417511,
1024
+ "learning_rate": 3.506941795528821e-05,
1025
+ "loss": 1.2576,
1026
+ "mean_token_accuracy": 0.7821117818355561,
1027
+ "num_tokens": 293519360.0,
1028
+ "step": 1120
1029
+ },
1030
+ {
1031
+ "epoch": 2.0,
1032
+ "eval_loss": 1.2395988702774048,
1033
+ "eval_mean_token_accuracy": 0.7735649704933166,
1034
+ "eval_num_tokens": 293961728.0,
1035
+ "eval_runtime": 2.8804,
1036
+ "eval_samples_per_second": 24.997,
1037
+ "eval_steps_per_second": 1.736,
1038
+ "step": 1122
1039
+ },
1040
+ {
1041
+ "epoch": 2.0142666072224698,
1042
+ "grad_norm": 0.6652095913887024,
1043
+ "learning_rate": 3.479846874098167e-05,
1044
+ "loss": 1.2741,
1045
+ "mean_token_accuracy": 0.7794669828353784,
1046
+ "num_tokens": 296058880.0,
1047
+ "step": 1130
1048
+ },
1049
+ {
1050
+ "epoch": 2.032099866250557,
1051
+ "grad_norm": 0.7581707835197449,
1052
+ "learning_rate": 3.452615277493544e-05,
1053
+ "loss": 1.199,
1054
+ "mean_token_accuracy": 0.7900658771395683,
1055
+ "num_tokens": 298680320.0,
1056
+ "step": 1140
1057
+ },
1058
+ {
1059
+ "epoch": 2.0499331252786446,
1060
+ "grad_norm": 0.5383137464523315,
1061
+ "learning_rate": 3.425250804148464e-05,
1062
+ "loss": 1.2304,
1063
+ "mean_token_accuracy": 0.7861943542957306,
1064
+ "num_tokens": 301301760.0,
1065
+ "step": 1150
1066
+ },
1067
+ {
1068
+ "epoch": 2.067766384306732,
1069
+ "grad_norm": 0.5952672958374023,
1070
+ "learning_rate": 3.3977572710309166e-05,
1071
+ "loss": 1.2445,
1072
+ "mean_token_accuracy": 0.7836834371089936,
1073
+ "num_tokens": 303923200.0,
1074
+ "step": 1160
1075
+ },
1076
+ {
1077
+ "epoch": 2.0855996433348194,
1078
+ "grad_norm": 0.646511435508728,
1079
+ "learning_rate": 3.3701385131109616e-05,
1080
+ "loss": 1.2019,
1081
+ "mean_token_accuracy": 0.7892128698527813,
1082
+ "num_tokens": 306544640.0,
1083
+ "step": 1170
1084
+ },
1085
+ {
1086
+ "epoch": 2.103432902362907,
1087
+ "grad_norm": 0.514665961265564,
1088
+ "learning_rate": 3.342398382825791e-05,
1089
+ "loss": 1.2614,
1090
+ "mean_token_accuracy": 0.7812080189585686,
1091
+ "num_tokens": 309166080.0,
1092
+ "step": 1180
1093
+ },
1094
+ {
1095
+ "epoch": 2.121266161390994,
1096
+ "grad_norm": 0.5490918159484863,
1097
+ "learning_rate": 3.31454074954238e-05,
1098
+ "loss": 1.2108,
1099
+ "mean_token_accuracy": 0.7876022845506668,
1100
+ "num_tokens": 311787520.0,
1101
+ "step": 1190
1102
+ },
1103
+ {
1104
+ "epoch": 2.1390994204190816,
1105
+ "grad_norm": 0.6549789309501648,
1106
+ "learning_rate": 3.286569499017752e-05,
1107
+ "loss": 1.2024,
1108
+ "mean_token_accuracy": 0.7898712314665317,
1109
+ "num_tokens": 314408960.0,
1110
+ "step": 1200
1111
+ },
1112
+ {
1113
+ "epoch": 2.156932679447169,
1114
+ "grad_norm": 0.5743107199668884,
1115
+ "learning_rate": 3.2584885328569785e-05,
1116
+ "loss": 1.1835,
1117
+ "mean_token_accuracy": 0.7928275369107723,
1118
+ "num_tokens": 317030400.0,
1119
+ "step": 1210
1120
+ },
1121
+ {
1122
+ "epoch": 2.1747659384752565,
1123
+ "grad_norm": 0.6201761960983276,
1124
+ "learning_rate": 3.2303017679689525e-05,
1125
+ "loss": 1.1939,
1126
+ "mean_token_accuracy": 0.7908047571778297,
1127
+ "num_tokens": 319651840.0,
1128
+ "step": 1220
1129
+ },
1130
+ {
1131
+ "epoch": 2.192599197503344,
1132
+ "grad_norm": 0.550830602645874,
1133
+ "learning_rate": 3.2020131360200354e-05,
1134
+ "loss": 1.1864,
1135
+ "mean_token_accuracy": 0.7930233284831048,
1136
+ "num_tokens": 322273280.0,
1137
+ "step": 1230
1138
+ },
1139
+ {
1140
+ "epoch": 2.2104324565314313,
1141
+ "grad_norm": 0.5972590446472168,
1142
+ "learning_rate": 3.173626582885645e-05,
1143
+ "loss": 1.2177,
1144
+ "mean_token_accuracy": 0.7885610058903694,
1145
+ "num_tokens": 324894720.0,
1146
+ "step": 1240
1147
+ },
1148
+ {
1149
+ "epoch": 2.2282657155595187,
1150
+ "grad_norm": 0.5069220066070557,
1151
+ "learning_rate": 3.145146068099858e-05,
1152
+ "loss": 1.2279,
1153
+ "mean_token_accuracy": 0.7862870991230011,
1154
+ "num_tokens": 327516160.0,
1155
+ "step": 1250
1156
+ },
1157
+ {
1158
+ "epoch": 2.246098974587606,
1159
+ "grad_norm": 0.5100940465927124,
1160
+ "learning_rate": 3.116575564303109e-05,
1161
+ "loss": 1.1777,
1162
+ "mean_token_accuracy": 0.7929233327507973,
1163
+ "num_tokens": 330137600.0,
1164
+ "step": 1260
1165
+ },
1166
+ {
1167
+ "epoch": 2.2639322336156935,
1168
+ "grad_norm": 0.6175646781921387,
1169
+ "learning_rate": 3.087919056688067e-05,
1170
+ "loss": 1.173,
1171
+ "mean_token_accuracy": 0.7950430497527122,
1172
+ "num_tokens": 332759040.0,
1173
+ "step": 1270
1174
+ },
1175
+ {
1176
+ "epoch": 2.2817654926437805,
1177
+ "grad_norm": 0.5933928489685059,
1178
+ "learning_rate": 3.059180542443746e-05,
1179
+ "loss": 1.2057,
1180
+ "mean_token_accuracy": 0.7882316343486309,
1181
+ "num_tokens": 335380480.0,
1182
+ "step": 1280
1183
+ },
1184
+ {
1185
+ "epoch": 2.299598751671868,
1186
+ "grad_norm": 0.4882568120956421,
1187
+ "learning_rate": 3.0303640301979635e-05,
1188
+ "loss": 1.1982,
1189
+ "mean_token_accuracy": 0.7903318881988526,
1190
+ "num_tokens": 338001920.0,
1191
+ "step": 1290
1192
+ },
1193
+ {
1194
+ "epoch": 2.3174320106999553,
1195
+ "grad_norm": 0.5135630369186401,
1196
+ "learning_rate": 3.0014735394581823e-05,
1197
+ "loss": 1.2045,
1198
+ "mean_token_accuracy": 0.7893250770866871,
1199
+ "num_tokens": 340623360.0,
1200
+ "step": 1300
1201
+ },
1202
+ {
1203
+ "epoch": 2.3352652697280427,
1204
+ "grad_norm": 0.5496548414230347,
1205
+ "learning_rate": 2.972513100050851e-05,
1206
+ "loss": 1.1883,
1207
+ "mean_token_accuracy": 0.791755847632885,
1208
+ "num_tokens": 343244800.0,
1209
+ "step": 1310
1210
+ },
1211
+ {
1212
+ "epoch": 2.35309852875613,
1213
+ "grad_norm": 0.5601444840431213,
1214
+ "learning_rate": 2.943486751559295e-05,
1215
+ "loss": 1.1263,
1216
+ "mean_token_accuracy": 0.7998171858489513,
1217
+ "num_tokens": 345866240.0,
1218
+ "step": 1320
1219
+ },
1220
+ {
1221
+ "epoch": 2.3709317877842175,
1222
+ "grad_norm": 0.5243608355522156,
1223
+ "learning_rate": 2.9143985427602516e-05,
1224
+ "loss": 1.2274,
1225
+ "mean_token_accuracy": 0.7866500526666641,
1226
+ "num_tokens": 348487680.0,
1227
+ "step": 1330
1228
+ },
1229
+ {
1230
+ "epoch": 2.388765046812305,
1231
+ "grad_norm": 0.573567807674408,
1232
+ "learning_rate": 2.88525253105912e-05,
1233
+ "loss": 1.1977,
1234
+ "mean_token_accuracy": 0.7901960171759128,
1235
+ "num_tokens": 351109120.0,
1236
+ "step": 1340
1237
+ },
1238
+ {
1239
+ "epoch": 2.4065983058403924,
1240
+ "grad_norm": 0.4823901951313019,
1241
+ "learning_rate": 2.856052781924012e-05,
1242
+ "loss": 1.1358,
1243
+ "mean_token_accuracy": 0.7998057343065739,
1244
+ "num_tokens": 353730560.0,
1245
+ "step": 1350
1246
+ },
1247
+ {
1248
+ "epoch": 2.4244315648684798,
1249
+ "grad_norm": 0.515941858291626,
1250
+ "learning_rate": 2.8268033683186697e-05,
1251
+ "loss": 1.1881,
1252
+ "mean_token_accuracy": 0.7909879550337792,
1253
+ "num_tokens": 356352000.0,
1254
+ "step": 1360
1255
+ },
1256
+ {
1257
+ "epoch": 2.442264823896567,
1258
+ "grad_norm": 0.5508431196212769,
1259
+ "learning_rate": 2.7975083701343497e-05,
1260
+ "loss": 1.171,
1261
+ "mean_token_accuracy": 0.7945770494639873,
1262
+ "num_tokens": 358973440.0,
1263
+ "step": 1370
1264
+ },
1265
+ {
1266
+ "epoch": 2.4600980829246546,
1267
+ "grad_norm": 0.5473787188529968,
1268
+ "learning_rate": 2.7681718736207298e-05,
1269
+ "loss": 1.1309,
1270
+ "mean_token_accuracy": 0.7995454497635365,
1271
+ "num_tokens": 361594880.0,
1272
+ "step": 1380
1273
+ },
1274
+ {
1275
+ "epoch": 2.477931341952742,
1276
+ "grad_norm": 0.6680080890655518,
1277
+ "learning_rate": 2.738797970815933e-05,
1278
+ "loss": 1.1848,
1279
+ "mean_token_accuracy": 0.7917638584971428,
1280
+ "num_tokens": 364216320.0,
1281
+ "step": 1390
1282
+ },
1283
+ {
1284
+ "epoch": 2.4957646009808294,
1285
+ "grad_norm": 0.49958300590515137,
1286
+ "learning_rate": 2.7093907589757468e-05,
1287
+ "loss": 1.1922,
1288
+ "mean_token_accuracy": 0.7899307683110237,
1289
+ "num_tokens": 366837760.0,
1290
+ "step": 1400
1291
+ },
1292
+ {
1293
+ "epoch": 2.5135978600089164,
1294
+ "grad_norm": 0.4601416289806366,
1295
+ "learning_rate": 2.6799543400021132e-05,
1296
+ "loss": 1.1448,
1297
+ "mean_token_accuracy": 0.7981104202568531,
1298
+ "num_tokens": 369459200.0,
1299
+ "step": 1410
1300
+ },
1301
+ {
1302
+ "epoch": 2.531431119037004,
1303
+ "grad_norm": 0.5619862079620361,
1304
+ "learning_rate": 2.650492819870965e-05,
1305
+ "loss": 1.1371,
1306
+ "mean_token_accuracy": 0.7994878143072128,
1307
+ "num_tokens": 372080640.0,
1308
+ "step": 1420
1309
+ },
1310
+ {
1311
+ "epoch": 2.549264378065091,
1312
+ "grad_norm": 0.6030861139297485,
1313
+ "learning_rate": 2.621010308059504e-05,
1314
+ "loss": 1.1695,
1315
+ "mean_token_accuracy": 0.7944411806762218,
1316
+ "num_tokens": 374702080.0,
1317
+ "step": 1430
1318
+ },
1319
+ {
1320
+ "epoch": 2.5670976370931786,
1321
+ "grad_norm": 0.5237954258918762,
1322
+ "learning_rate": 2.5915109169729814e-05,
1323
+ "loss": 1.1372,
1324
+ "mean_token_accuracy": 0.7982218623161316,
1325
+ "num_tokens": 377323520.0,
1326
+ "step": 1440
1327
+ },
1328
+ {
1329
+ "epoch": 2.584930896121266,
1330
+ "grad_norm": 0.5144650936126709,
1331
+ "learning_rate": 2.5619987613710756e-05,
1332
+ "loss": 1.1896,
1333
+ "mean_token_accuracy": 0.7915455557405948,
1334
+ "num_tokens": 379944960.0,
1335
+ "step": 1450
1336
+ },
1337
+ {
1338
+ "epoch": 2.6027641551493534,
1339
+ "grad_norm": 0.5404797196388245,
1340
+ "learning_rate": 2.5324779577939394e-05,
1341
+ "loss": 1.1364,
1342
+ "mean_token_accuracy": 0.7987703062593937,
1343
+ "num_tokens": 382566400.0,
1344
+ "step": 1460
1345
+ },
1346
+ {
1347
+ "epoch": 2.620597414177441,
1348
+ "grad_norm": 0.4770708382129669,
1349
+ "learning_rate": 2.5029526239880015e-05,
1350
+ "loss": 1.1541,
1351
+ "mean_token_accuracy": 0.7973127588629723,
1352
+ "num_tokens": 385187840.0,
1353
+ "step": 1470
1354
+ },
1355
+ {
1356
+ "epoch": 2.6384306732055283,
1357
+ "grad_norm": 0.5579174757003784,
1358
+ "learning_rate": 2.473426878331591e-05,
1359
+ "loss": 1.1695,
1360
+ "mean_token_accuracy": 0.7945644512772561,
1361
+ "num_tokens": 387809280.0,
1362
+ "step": 1480
1363
+ },
1364
+ {
1365
+ "epoch": 2.6562639322336157,
1366
+ "grad_norm": 0.5287127494812012,
1367
+ "learning_rate": 2.443904839260488e-05,
1368
+ "loss": 1.1615,
1369
+ "mean_token_accuracy": 0.7949182473123073,
1370
+ "num_tokens": 390430720.0,
1371
+ "step": 1490
1372
+ },
1373
+ {
1374
+ "epoch": 2.674097191261703,
1375
+ "grad_norm": 0.5106226801872253,
1376
+ "learning_rate": 2.4143906246934537e-05,
1377
+ "loss": 1.1201,
1378
+ "mean_token_accuracy": 0.8007564425468445,
1379
+ "num_tokens": 393052160.0,
1380
+ "step": 1500
1381
+ },
1382
+ {
1383
+ "epoch": 2.6919304502897905,
1384
+ "grad_norm": 0.48907479643821716,
1385
+ "learning_rate": 2.384888351457838e-05,
1386
+ "loss": 1.1908,
1387
+ "mean_token_accuracy": 0.7911803111433983,
1388
+ "num_tokens": 395673600.0,
1389
+ "step": 1510
1390
+ },
1391
+ {
1392
+ "epoch": 2.709763709317878,
1393
+ "grad_norm": 0.5045855045318604,
1394
+ "learning_rate": 2.3554021347153403e-05,
1395
+ "loss": 1.1483,
1396
+ "mean_token_accuracy": 0.7968017213046551,
1397
+ "num_tokens": 398295040.0,
1398
+ "step": 1520
1399
+ },
1400
+ {
1401
+ "epoch": 2.7275969683459653,
1402
+ "grad_norm": 0.505662739276886,
1403
+ "learning_rate": 2.3259360873879992e-05,
1404
+ "loss": 1.1749,
1405
+ "mean_token_accuracy": 0.7930756144225597,
1406
+ "num_tokens": 400916480.0,
1407
+ "step": 1530
1408
+ },
1409
+ {
1410
+ "epoch": 2.7454302273740527,
1411
+ "grad_norm": 0.5283376574516296,
1412
+ "learning_rate": 2.2964943195844966e-05,
1413
+ "loss": 1.173,
1414
+ "mean_token_accuracy": 0.7921638391911984,
1415
+ "num_tokens": 403537920.0,
1416
+ "step": 1540
1417
+ },
1418
+ {
1419
+ "epoch": 2.76326348640214,
1420
+ "grad_norm": 0.5634790062904358,
1421
+ "learning_rate": 2.2670809380268548e-05,
1422
+ "loss": 1.1779,
1423
+ "mean_token_accuracy": 0.7936561174690724,
1424
+ "num_tokens": 406159360.0,
1425
+ "step": 1550
1426
+ },
1427
+ {
1428
+ "epoch": 2.7810967454302276,
1429
+ "grad_norm": 0.48228803277015686,
1430
+ "learning_rate": 2.237700045477607e-05,
1431
+ "loss": 1.14,
1432
+ "mean_token_accuracy": 0.796902097016573,
1433
+ "num_tokens": 408780800.0,
1434
+ "step": 1560
1435
+ },
1436
+ {
1437
+ "epoch": 2.798930004458315,
1438
+ "grad_norm": 0.4753395617008209,
1439
+ "learning_rate": 2.2083557401675115e-05,
1440
+ "loss": 1.1294,
1441
+ "mean_token_accuracy": 0.7987142026424408,
1442
+ "num_tokens": 411402240.0,
1443
+ "step": 1570
1444
+ },
1445
+ {
1446
+ "epoch": 2.8167632634864024,
1447
+ "grad_norm": 0.5348941683769226,
1448
+ "learning_rate": 2.1790521152239137e-05,
1449
+ "loss": 1.1581,
1450
+ "mean_token_accuracy": 0.7944865994155407,
1451
+ "num_tokens": 414023680.0,
1452
+ "step": 1580
1453
+ },
1454
+ {
1455
+ "epoch": 2.83459652251449,
1456
+ "grad_norm": 0.5610408782958984,
1457
+ "learning_rate": 2.1497932580998053e-05,
1458
+ "loss": 1.1609,
1459
+ "mean_token_accuracy": 0.795001070201397,
1460
+ "num_tokens": 416645120.0,
1461
+ "step": 1590
1462
+ },
1463
+ {
1464
+ "epoch": 2.8524297815425768,
1465
+ "grad_norm": 0.5431333184242249,
1466
+ "learning_rate": 2.1205832500036838e-05,
1467
+ "loss": 1.1405,
1468
+ "mean_token_accuracy": 0.7983539171516896,
1469
+ "num_tokens": 419266560.0,
1470
+ "step": 1600
1471
+ },
1472
+ {
1473
+ "epoch": 2.870263040570664,
1474
+ "grad_norm": 0.5650882720947266,
1475
+ "learning_rate": 2.0914261653302747e-05,
1476
+ "loss": 1.17,
1477
+ "mean_token_accuracy": 0.7925355695188046,
1478
+ "num_tokens": 421888000.0,
1479
+ "step": 1610
1480
+ },
1481
+ {
1482
+ "epoch": 2.8880962995987516,
1483
+ "grad_norm": 0.5435991287231445,
1484
+ "learning_rate": 2.062326071092218e-05,
1485
+ "loss": 1.1562,
1486
+ "mean_token_accuracy": 0.7965150982141495,
1487
+ "num_tokens": 424509440.0,
1488
+ "step": 1620
1489
+ },
1490
+ {
1491
+ "epoch": 2.905929558626839,
1492
+ "grad_norm": 0.4466145932674408,
1493
+ "learning_rate": 2.0332870263527675e-05,
1494
+ "loss": 1.1199,
1495
+ "mean_token_accuracy": 0.8016460791230202,
1496
+ "num_tokens": 427130880.0,
1497
+ "step": 1630
1498
+ },
1499
+ {
1500
+ "epoch": 2.9237628176549264,
1501
+ "grad_norm": 0.5083450675010681,
1502
+ "learning_rate": 2.0043130816596136e-05,
1503
+ "loss": 1.1467,
1504
+ "mean_token_accuracy": 0.7959990955889225,
1505
+ "num_tokens": 429752320.0,
1506
+ "step": 1640
1507
+ },
1508
+ {
1509
+ "epoch": 2.941596076683014,
1510
+ "grad_norm": 0.49085912108421326,
1511
+ "learning_rate": 1.9754082784798843e-05,
1512
+ "loss": 1.1565,
1513
+ "mean_token_accuracy": 0.7959273479878902,
1514
+ "num_tokens": 432373760.0,
1515
+ "step": 1650
1516
+ },
1517
+ {
1518
+ "epoch": 2.9594293357111012,
1519
+ "grad_norm": 0.48444733023643494,
1520
+ "learning_rate": 1.9465766486364143e-05,
1521
+ "loss": 1.1342,
1522
+ "mean_token_accuracy": 0.797958905249834,
1523
+ "num_tokens": 434995200.0,
1524
+ "step": 1660
1525
+ },
1526
+ {
1527
+ "epoch": 2.9772625947391886,
1528
+ "grad_norm": 0.5343450903892517,
1529
+ "learning_rate": 1.9178222137453656e-05,
1530
+ "loss": 1.141,
1531
+ "mean_token_accuracy": 0.7985142156481743,
1532
+ "num_tokens": 437616640.0,
1533
+ "step": 1670
1534
+ },
1535
+ {
1536
+ "epoch": 2.995095853767276,
1537
+ "grad_norm": 0.5583868026733398,
1538
+ "learning_rate": 1.8891489846552646e-05,
1539
+ "loss": 1.1432,
1540
+ "mean_token_accuracy": 0.7971971161663532,
1541
+ "num_tokens": 440238080.0,
1542
+ "step": 1680
1543
+ },
1544
+ {
1545
+ "epoch": 3.0,
1546
+ "eval_loss": 1.1234296560287476,
1547
+ "eval_mean_token_accuracy": 0.7894601821899414,
1548
+ "eval_num_tokens": 440942592.0,
1549
+ "eval_runtime": 2.9179,
1550
+ "eval_samples_per_second": 24.675,
1551
+ "eval_steps_per_second": 1.714,
1552
+ "step": 1683
1553
+ },
1554
+ {
1555
+ "epoch": 3.0124832813196614,
1556
+ "grad_norm": 0.5347409844398499,
1557
+ "learning_rate": 1.860560960887546e-05,
1558
+ "loss": 1.1253,
1559
+ "mean_token_accuracy": 0.8009015726737487,
1560
+ "num_tokens": 442777600.0,
1561
+ "step": 1690
1562
+ },
1563
+ {
1564
+ "epoch": 3.0303165403477483,
1565
+ "grad_norm": 0.5598998069763184,
1566
+ "learning_rate": 1.8320621300786688e-05,
1567
+ "loss": 1.106,
1568
+ "mean_token_accuracy": 0.8023651242256165,
1569
+ "num_tokens": 445399040.0,
1570
+ "step": 1700
1571
+ },
1572
+ {
1573
+ "epoch": 3.0481497993758357,
1574
+ "grad_norm": 0.5113139152526855,
1575
+ "learning_rate": 1.8036564674239043e-05,
1576
+ "loss": 1.0887,
1577
+ "mean_token_accuracy": 0.8040673099458218,
1578
+ "num_tokens": 448020480.0,
1579
+ "step": 1710
1580
+ },
1581
+ {
1582
+ "epoch": 3.065983058403923,
1583
+ "grad_norm": 0.44441869854927063,
1584
+ "learning_rate": 1.7753479351228444e-05,
1585
+ "loss": 1.1185,
1586
+ "mean_token_accuracy": 0.7999229028820991,
1587
+ "num_tokens": 450641920.0,
1588
+ "step": 1720
1589
+ },
1590
+ {
1591
+ "epoch": 3.0838163174320106,
1592
+ "grad_norm": 0.47339409589767456,
1593
+ "learning_rate": 1.7471404818267316e-05,
1594
+ "loss": 1.0925,
1595
+ "mean_token_accuracy": 0.8034070417284965,
1596
+ "num_tokens": 453263360.0,
1597
+ "step": 1730
1598
+ },
1599
+ {
1600
+ "epoch": 3.101649576460098,
1601
+ "grad_norm": 0.47856712341308594,
1602
+ "learning_rate": 1.7190380420876766e-05,
1603
+ "loss": 1.0911,
1604
+ "mean_token_accuracy": 0.8045188069343567,
1605
+ "num_tokens": 455884800.0,
1606
+ "step": 1740
1607
+ },
1608
+ {
1609
+ "epoch": 3.1194828354881854,
1610
+ "grad_norm": 0.4918275773525238,
1611
+ "learning_rate": 1.6910445358098415e-05,
1612
+ "loss": 1.1109,
1613
+ "mean_token_accuracy": 0.8015380732715129,
1614
+ "num_tokens": 458506240.0,
1615
+ "step": 1750
1616
+ },
1617
+ {
1618
+ "epoch": 3.137316094516273,
1619
+ "grad_norm": 0.5043804049491882,
1620
+ "learning_rate": 1.6631638677026674e-05,
1621
+ "loss": 1.1364,
1622
+ "mean_token_accuracy": 0.7974261090159416,
1623
+ "num_tokens": 461127680.0,
1624
+ "step": 1760
1625
+ },
1626
+ {
1627
+ "epoch": 3.15514935354436,
1628
+ "grad_norm": 0.4904772937297821,
1629
+ "learning_rate": 1.6353999267362237e-05,
1630
+ "loss": 1.0838,
1631
+ "mean_token_accuracy": 0.806386636942625,
1632
+ "num_tokens": 463749120.0,
1633
+ "step": 1770
1634
+ },
1635
+ {
1636
+ "epoch": 3.1729826125724476,
1637
+ "grad_norm": 0.48294153809547424,
1638
+ "learning_rate": 1.607756585598743e-05,
1639
+ "loss": 1.1389,
1640
+ "mean_token_accuracy": 0.7980314172804356,
1641
+ "num_tokens": 466370560.0,
1642
+ "step": 1780
1643
+ },
1644
+ {
1645
+ "epoch": 3.190815871600535,
1646
+ "grad_norm": 0.5139598250389099,
1647
+ "learning_rate": 1.580237700156443e-05,
1648
+ "loss": 1.1095,
1649
+ "mean_token_accuracy": 0.8025780878961086,
1650
+ "num_tokens": 468992000.0,
1651
+ "step": 1790
1652
+ },
1653
+ {
1654
+ "epoch": 3.2086491306286224,
1655
+ "grad_norm": 0.47298020124435425,
1656
+ "learning_rate": 1.5528471089156804e-05,
1657
+ "loss": 1.0931,
1658
+ "mean_token_accuracy": 0.8041585251688957,
1659
+ "num_tokens": 471613440.0,
1660
+ "step": 1800
1661
+ },
1662
+ {
1663
+ "epoch": 3.22648238965671,
1664
+ "grad_norm": 0.41903576254844666,
1665
+ "learning_rate": 1.5255886324875363e-05,
1666
+ "loss": 1.0954,
1667
+ "mean_token_accuracy": 0.8047371134161949,
1668
+ "num_tokens": 474234880.0,
1669
+ "step": 1810
1670
+ },
1671
+ {
1672
+ "epoch": 3.2443156486847973,
1673
+ "grad_norm": 0.4610045552253723,
1674
+ "learning_rate": 1.4984660730548856e-05,
1675
+ "loss": 1.1095,
1676
+ "mean_token_accuracy": 0.8022483348846435,
1677
+ "num_tokens": 476856320.0,
1678
+ "step": 1820
1679
+ },
1680
+ {
1681
+ "epoch": 3.2621489077128847,
1682
+ "grad_norm": 0.5105413198471069,
1683
+ "learning_rate": 1.4714832138420553e-05,
1684
+ "loss": 1.1258,
1685
+ "mean_token_accuracy": 0.7983997173607349,
1686
+ "num_tokens": 479477760.0,
1687
+ "step": 1830
1688
+ },
1689
+ {
1690
+ "epoch": 3.279982166740972,
1691
+ "grad_norm": 0.4393883943557739,
1692
+ "learning_rate": 1.4446438185871092e-05,
1693
+ "loss": 1.0566,
1694
+ "mean_token_accuracy": 0.8093036323785782,
1695
+ "num_tokens": 482099200.0,
1696
+ "step": 1840
1697
+ },
1698
+ {
1699
+ "epoch": 3.297815425769059,
1700
+ "grad_norm": 0.4618268311023712,
1701
+ "learning_rate": 1.4179516310168623e-05,
1702
+ "loss": 1.0697,
1703
+ "mean_token_accuracy": 0.8082189694046974,
1704
+ "num_tokens": 484720640.0,
1705
+ "step": 1850
1706
+ },
1707
+ {
1708
+ "epoch": 3.3156486847971465,
1709
+ "grad_norm": 0.47192856669425964,
1710
+ "learning_rate": 1.3914103743246757e-05,
1711
+ "loss": 1.1203,
1712
+ "mean_token_accuracy": 0.8002961635589599,
1713
+ "num_tokens": 487342080.0,
1714
+ "step": 1860
1715
+ },
1716
+ {
1717
+ "epoch": 3.333481943825234,
1718
+ "grad_norm": 0.4617849290370941,
1719
+ "learning_rate": 1.3650237506511331e-05,
1720
+ "loss": 1.0664,
1721
+ "mean_token_accuracy": 0.8089914366602897,
1722
+ "num_tokens": 489963520.0,
1723
+ "step": 1870
1724
+ },
1725
+ {
1726
+ "epoch": 3.3513152028533213,
1727
+ "grad_norm": 0.46263769268989563,
1728
+ "learning_rate": 1.3387954405676329e-05,
1729
+ "loss": 1.0855,
1730
+ "mean_token_accuracy": 0.8045512497425079,
1731
+ "num_tokens": 492584960.0,
1732
+ "step": 1880
1733
+ },
1734
+ {
1735
+ "epoch": 3.3691484618814087,
1736
+ "grad_norm": 0.47842440009117126,
1737
+ "learning_rate": 1.3127291025630073e-05,
1738
+ "loss": 1.0705,
1739
+ "mean_token_accuracy": 0.8085929870605468,
1740
+ "num_tokens": 495206400.0,
1741
+ "step": 1890
1742
+ },
1743
+ {
1744
+ "epoch": 3.386981720909496,
1745
+ "grad_norm": 0.43500861525535583,
1746
+ "learning_rate": 1.2868283725332028e-05,
1747
+ "loss": 1.0635,
1748
+ "mean_token_accuracy": 0.8091395162045956,
1749
+ "num_tokens": 497827840.0,
1750
+ "step": 1900
1751
+ },
1752
+ {
1753
+ "epoch": 3.4048149799375835,
1754
+ "grad_norm": 0.508797287940979,
1755
+ "learning_rate": 1.261096863274136e-05,
1756
+ "loss": 1.1052,
1757
+ "mean_token_accuracy": 0.8021300226449967,
1758
+ "num_tokens": 500449280.0,
1759
+ "step": 1910
1760
+ },
1761
+ {
1762
+ "epoch": 3.422648238965671,
1763
+ "grad_norm": 0.42378219962120056,
1764
+ "learning_rate": 1.2355381639777472e-05,
1765
+ "loss": 1.0936,
1766
+ "mean_token_accuracy": 0.8043497301638126,
1767
+ "num_tokens": 503070720.0,
1768
+ "step": 1920
1769
+ },
1770
+ {
1771
+ "epoch": 3.4404814979937584,
1772
+ "grad_norm": 0.4547331631183624,
1773
+ "learning_rate": 1.2101558397313634e-05,
1774
+ "loss": 1.1142,
1775
+ "mean_token_accuracy": 0.8015869230031967,
1776
+ "num_tokens": 505692160.0,
1777
+ "step": 1930
1778
+ },
1779
+ {
1780
+ "epoch": 3.4583147570218458,
1781
+ "grad_norm": 0.4319451153278351,
1782
+ "learning_rate": 1.1849534310204152e-05,
1783
+ "loss": 1.0835,
1784
+ "mean_token_accuracy": 0.8058400988578797,
1785
+ "num_tokens": 508313600.0,
1786
+ "step": 1940
1787
+ },
1788
+ {
1789
+ "epoch": 3.476148016049933,
1790
+ "grad_norm": 0.4666835069656372,
1791
+ "learning_rate": 1.1599344532345877e-05,
1792
+ "loss": 1.1072,
1793
+ "mean_token_accuracy": 0.801591893285513,
1794
+ "num_tokens": 510935040.0,
1795
+ "step": 1950
1796
+ },
1797
+ {
1798
+ "epoch": 3.4939812750780206,
1799
+ "grad_norm": 0.47252121567726135,
1800
+ "learning_rate": 1.1351023961774749e-05,
1801
+ "loss": 1.1245,
1802
+ "mean_token_accuracy": 0.7985180281102657,
1803
+ "num_tokens": 513556480.0,
1804
+ "step": 1960
1805
+ }
1806
+ ],
1807
+ "logging_steps": 10,
1808
+ "max_steps": 2800,
1809
+ "num_input_tokens_seen": 0,
1810
+ "num_train_epochs": 5,
1811
+ "save_steps": 140,
1812
+ "stateful_callbacks": {
1813
+ "TrainerControl": {
1814
+ "args": {
1815
+ "should_epoch_stop": false,
1816
+ "should_evaluate": false,
1817
+ "should_log": false,
1818
+ "should_save": true,
1819
+ "should_training_stop": false
1820
+ },
1821
+ "attributes": {}
1822
+ }
1823
+ },
1824
+ "total_flos": 2.557045311591678e+18,
1825
+ "train_batch_size": 16,
1826
+ "trial_name": null,
1827
+ "trial_params": null
1828
+ }
checkpoint-1960/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7409b037df1da3374237680bbedeb80953c01e549dccfb71a127e04d8bc7415b
3
+ size 7352
checkpoint-2100/config.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_configs": [
3
+ {
4
+ "vocab_size": 49152,
5
+ "max_position_embeddings": 8192,
6
+ "hidden_size": 960,
7
+ "intermediate_size": 2560,
8
+ "num_hidden_layers": 32,
9
+ "num_attention_heads": 15,
10
+ "num_key_value_heads": 5,
11
+ "hidden_act": "silu",
12
+ "initializer_range": 0.02,
13
+ "rms_norm_eps": 1e-05,
14
+ "pretraining_tp": 1,
15
+ "use_cache": true,
16
+ "rope_theta": 100000,
17
+ "rope_scaling": null,
18
+ "attention_bias": false,
19
+ "attention_dropout": 0.0,
20
+ "mlp_bias": false,
21
+ "head_dim": 64,
22
+ "return_dict": true,
23
+ "output_hidden_states": false,
24
+ "output_attentions": false,
25
+ "torchscript": false,
26
+ "torch_dtype": "float32",
27
+ "use_bfloat16": false,
28
+ "tf_legacy_loss": false,
29
+ "pruned_heads": {},
30
+ "tie_word_embeddings": true,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "is_decoder": false,
34
+ "cross_attention_hidden_size": null,
35
+ "add_cross_attention": false,
36
+ "tie_encoder_decoder": false,
37
+ "max_length": 20,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "early_stopping": false,
41
+ "num_beams": 1,
42
+ "num_beam_groups": 1,
43
+ "diversity_penalty": 0.0,
44
+ "temperature": 1.0,
45
+ "top_k": 50,
46
+ "top_p": 1.0,
47
+ "typical_p": 1.0,
48
+ "repetition_penalty": 1.0,
49
+ "length_penalty": 1.0,
50
+ "no_repeat_ngram_size": 0,
51
+ "encoder_no_repeat_ngram_size": 0,
52
+ "bad_words_ids": null,
53
+ "num_return_sequences": 1,
54
+ "output_scores": false,
55
+ "return_dict_in_generate": false,
56
+ "forced_bos_token_id": null,
57
+ "forced_eos_token_id": null,
58
+ "remove_invalid_values": false,
59
+ "exponential_decay_length_penalty": null,
60
+ "suppress_tokens": null,
61
+ "begin_suppress_tokens": null,
62
+ "architectures": [
63
+ "LlamaForCausalLM"
64
+ ],
65
+ "finetuning_task": null,
66
+ "id2label": {
67
+ "0": "LABEL_0",
68
+ "1": "LABEL_1"
69
+ },
70
+ "label2id": {
71
+ "LABEL_0": 0,
72
+ "LABEL_1": 1
73
+ },
74
+ "tokenizer_class": null,
75
+ "prefix": null,
76
+ "bos_token_id": 1,
77
+ "pad_token_id": 2,
78
+ "eos_token_id": 2,
79
+ "sep_token_id": null,
80
+ "decoder_start_token_id": null,
81
+ "task_specific_params": null,
82
+ "problem_type": null,
83
+ "_name_or_path": "HuggingFaceTB/SmolLM2-360M-Instruct",
84
+ "_attn_implementation_autoset": true,
85
+ "transformers_version": "4.51.3",
86
+ "is_llama_config": true,
87
+ "model_type": "llama",
88
+ "rope_interleaved": false,
89
+ "transformers.js_config": {
90
+ "kv_cache_dtype": {
91
+ "q4f16": "float16",
92
+ "fp16": "float16"
93
+ }
94
+ }
95
+ },
96
+ {
97
+ "vocab_size": 49152,
98
+ "max_position_embeddings": 8192,
99
+ "hidden_size": 576,
100
+ "intermediate_size": 1536,
101
+ "num_hidden_layers": 30,
102
+ "num_attention_heads": 9,
103
+ "num_key_value_heads": 3,
104
+ "hidden_act": "silu",
105
+ "initializer_range": 0.041666666666666664,
106
+ "rms_norm_eps": 1e-05,
107
+ "pretraining_tp": 1,
108
+ "use_cache": true,
109
+ "rope_theta": 100000,
110
+ "rope_scaling": null,
111
+ "attention_bias": false,
112
+ "attention_dropout": 0.0,
113
+ "mlp_bias": false,
114
+ "head_dim": 64,
115
+ "return_dict": true,
116
+ "output_hidden_states": false,
117
+ "output_attentions": false,
118
+ "torchscript": false,
119
+ "torch_dtype": "float32",
120
+ "use_bfloat16": false,
121
+ "tf_legacy_loss": false,
122
+ "pruned_heads": {},
123
+ "tie_word_embeddings": true,
124
+ "chunk_size_feed_forward": 0,
125
+ "is_encoder_decoder": false,
126
+ "is_decoder": false,
127
+ "cross_attention_hidden_size": null,
128
+ "add_cross_attention": false,
129
+ "tie_encoder_decoder": false,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "do_sample": false,
133
+ "early_stopping": false,
134
+ "num_beams": 1,
135
+ "num_beam_groups": 1,
136
+ "diversity_penalty": 0.0,
137
+ "temperature": 1.0,
138
+ "top_k": 50,
139
+ "top_p": 1.0,
140
+ "typical_p": 1.0,
141
+ "repetition_penalty": 1.0,
142
+ "length_penalty": 1.0,
143
+ "no_repeat_ngram_size": 0,
144
+ "encoder_no_repeat_ngram_size": 0,
145
+ "bad_words_ids": null,
146
+ "num_return_sequences": 1,
147
+ "output_scores": false,
148
+ "return_dict_in_generate": false,
149
+ "forced_bos_token_id": null,
150
+ "forced_eos_token_id": null,
151
+ "remove_invalid_values": false,
152
+ "exponential_decay_length_penalty": null,
153
+ "suppress_tokens": null,
154
+ "begin_suppress_tokens": null,
155
+ "architectures": [
156
+ "LlamaForCausalLM"
157
+ ],
158
+ "finetuning_task": null,
159
+ "id2label": {
160
+ "0": "LABEL_0",
161
+ "1": "LABEL_1"
162
+ },
163
+ "label2id": {
164
+ "LABEL_0": 0,
165
+ "LABEL_1": 1
166
+ },
167
+ "tokenizer_class": null,
168
+ "prefix": null,
169
+ "bos_token_id": 1,
170
+ "pad_token_id": 2,
171
+ "eos_token_id": 2,
172
+ "sep_token_id": null,
173
+ "decoder_start_token_id": null,
174
+ "task_specific_params": null,
175
+ "problem_type": null,
176
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct",
177
+ "_attn_implementation_autoset": true,
178
+ "transformers_version": "4.51.3",
179
+ "is_llama_config": true,
180
+ "model_type": "llama",
181
+ "rope_interleaved": false,
182
+ "transformers.js_config": {
183
+ "kv_cache_dtype": {
184
+ "q4f16": "float16",
185
+ "fp16": "float16"
186
+ }
187
+ }
188
+ }
189
+ ],
190
+ "num_models": 2
191
+ }
checkpoint-2100/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2100/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10728850487aa0029e42533bc3fe6b7d52368a9ff25282628028404d7a3bc2af
3
+ size 2362909008
checkpoint-2100/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7d0e428eaaebbd39500df7b70c85e0c423ad7178fcc03acaa80860c51ef011a
3
+ size 14244
checkpoint-2100/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06fa43f3b2822cbe09e033c883f8cf878355d7c1ed07e07eb0e0f866af29d09c
3
+ size 1064
checkpoint-2100/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
checkpoint-2100/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2100/tokenizer_config.json ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<repo_name>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "4": {
37
+ "content": "<reponame>",
38
+ "lstrip": false,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "5": {
45
+ "content": "<file_sep>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "6": {
53
+ "content": "<filename>",
54
+ "lstrip": false,
55
+ "normalized": false,
56
+ "rstrip": false,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "7": {
61
+ "content": "<gh_stars>",
62
+ "lstrip": false,
63
+ "normalized": false,
64
+ "rstrip": false,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "8": {
69
+ "content": "<issue_start>",
70
+ "lstrip": false,
71
+ "normalized": false,
72
+ "rstrip": false,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "9": {
77
+ "content": "<issue_comment>",
78
+ "lstrip": false,
79
+ "normalized": false,
80
+ "rstrip": false,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "10": {
85
+ "content": "<issue_closed>",
86
+ "lstrip": false,
87
+ "normalized": false,
88
+ "rstrip": false,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "11": {
93
+ "content": "<jupyter_start>",
94
+ "lstrip": false,
95
+ "normalized": false,
96
+ "rstrip": false,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "12": {
101
+ "content": "<jupyter_text>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "13": {
109
+ "content": "<jupyter_code>",
110
+ "lstrip": false,
111
+ "normalized": false,
112
+ "rstrip": false,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "14": {
117
+ "content": "<jupyter_output>",
118
+ "lstrip": false,
119
+ "normalized": false,
120
+ "rstrip": false,
121
+ "single_word": false,
122
+ "special": true
123
+ },
124
+ "15": {
125
+ "content": "<jupyter_script>",
126
+ "lstrip": false,
127
+ "normalized": false,
128
+ "rstrip": false,
129
+ "single_word": false,
130
+ "special": true
131
+ },
132
+ "16": {
133
+ "content": "<empty_output>",
134
+ "lstrip": false,
135
+ "normalized": false,
136
+ "rstrip": false,
137
+ "single_word": false,
138
+ "special": true
139
+ }
140
+ },
141
+ "additional_special_tokens": [
142
+ "<|im_start|>",
143
+ "<|im_end|>"
144
+ ],
145
+ "bos_token": "<|im_start|>",
146
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
147
+ "clean_up_tokenization_spaces": false,
148
+ "eos_token": "<|im_end|>",
149
+ "extra_special_tokens": {},
150
+ "model_max_length": 8192,
151
+ "pad_token": "<|im_end|>",
152
+ "tokenizer_class": "GPT2Tokenizer",
153
+ "unk_token": "<|endoftext|>",
154
+ "vocab_size": 49152
155
+ }
checkpoint-2100/trainer_state.json ADDED
@@ -0,0 +1,1954 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.743646901471244,
6
+ "eval_steps": 500,
7
+ "global_step": 2100,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017833259028087384,
14
+ "grad_norm": 33.452999114990234,
15
+ "learning_rate": 3.2142857142857143e-06,
16
+ "loss": 8.9944,
17
+ "mean_token_accuracy": 0.32701094649964946,
18
+ "num_tokens": 2621440.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.03566651805617477,
23
+ "grad_norm": 2.596579074859619,
24
+ "learning_rate": 6.785714285714285e-06,
25
+ "loss": 4.9649,
26
+ "mean_token_accuracy": 0.5848119974136352,
27
+ "num_tokens": 5242880.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.05349977708426215,
32
+ "grad_norm": 1.8979898691177368,
33
+ "learning_rate": 1.0357142857142859e-05,
34
+ "loss": 4.1262,
35
+ "mean_token_accuracy": 0.5906830102205276,
36
+ "num_tokens": 7864320.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.07133303611234953,
41
+ "grad_norm": 1.1842058897018433,
42
+ "learning_rate": 1.392857142857143e-05,
43
+ "loss": 3.7327,
44
+ "mean_token_accuracy": 0.608327355235815,
45
+ "num_tokens": 10485760.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.08916629514043692,
50
+ "grad_norm": 0.922074556350708,
51
+ "learning_rate": 1.75e-05,
52
+ "loss": 3.4492,
53
+ "mean_token_accuracy": 0.6281662203371525,
54
+ "num_tokens": 13107200.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.1069995541685243,
59
+ "grad_norm": 0.7813541889190674,
60
+ "learning_rate": 2.107142857142857e-05,
61
+ "loss": 3.1416,
62
+ "mean_token_accuracy": 0.648241326212883,
63
+ "num_tokens": 15728640.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.12483281319661169,
68
+ "grad_norm": 1.1468125581741333,
69
+ "learning_rate": 2.4642857142857145e-05,
70
+ "loss": 2.8778,
71
+ "mean_token_accuracy": 0.6583510167896748,
72
+ "num_tokens": 18350080.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.14266607222469907,
77
+ "grad_norm": 1.0490431785583496,
78
+ "learning_rate": 2.8214285714285714e-05,
79
+ "loss": 2.7108,
80
+ "mean_token_accuracy": 0.663086223602295,
81
+ "num_tokens": 20971520.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.16049933125278645,
86
+ "grad_norm": 0.7813755869865417,
87
+ "learning_rate": 3.1785714285714284e-05,
88
+ "loss": 2.6039,
89
+ "mean_token_accuracy": 0.6650082431733608,
90
+ "num_tokens": 23592960.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.17833259028087384,
95
+ "grad_norm": 0.6500654220581055,
96
+ "learning_rate": 3.5357142857142864e-05,
97
+ "loss": 2.4513,
98
+ "mean_token_accuracy": 0.674950385093689,
99
+ "num_tokens": 26214400.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.19616584930896122,
104
+ "grad_norm": 0.81620192527771,
105
+ "learning_rate": 3.892857142857143e-05,
106
+ "loss": 2.3602,
107
+ "mean_token_accuracy": 0.6777540303766727,
108
+ "num_tokens": 28835840.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.2139991083370486,
113
+ "grad_norm": 0.6131238341331482,
114
+ "learning_rate": 4.25e-05,
115
+ "loss": 2.3545,
116
+ "mean_token_accuracy": 0.6758881151676178,
117
+ "num_tokens": 31457280.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.231832367365136,
122
+ "grad_norm": 0.9951689839363098,
123
+ "learning_rate": 4.607142857142857e-05,
124
+ "loss": 2.2624,
125
+ "mean_token_accuracy": 0.6806462205946445,
126
+ "num_tokens": 34078720.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.24966562639322337,
131
+ "grad_norm": 0.7796569466590881,
132
+ "learning_rate": 4.964285714285715e-05,
133
+ "loss": 2.1505,
134
+ "mean_token_accuracy": 0.6918394304811954,
135
+ "num_tokens": 36700160.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.26749888542131073,
140
+ "grad_norm": 0.8614610433578491,
141
+ "learning_rate": 4.9998587698517684e-05,
142
+ "loss": 2.1591,
143
+ "mean_token_accuracy": 0.68697941750288,
144
+ "num_tokens": 39321600.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.28533214444939814,
149
+ "grad_norm": 0.6178464889526367,
150
+ "learning_rate": 4.9993705873562665e-05,
151
+ "loss": 2.1048,
152
+ "mean_token_accuracy": 0.6911555036902428,
153
+ "num_tokens": 41943040.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.3031654034774855,
158
+ "grad_norm": 0.7255717515945435,
159
+ "learning_rate": 4.998533777009954e-05,
160
+ "loss": 2.0847,
161
+ "mean_token_accuracy": 0.6920825429260731,
162
+ "num_tokens": 44564480.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.3209986625055729,
167
+ "grad_norm": 0.7800336480140686,
168
+ "learning_rate": 4.9973484555363726e-05,
169
+ "loss": 2.049,
170
+ "mean_token_accuracy": 0.6982546880841255,
171
+ "num_tokens": 47185920.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.33883192153366026,
176
+ "grad_norm": 0.886745274066925,
177
+ "learning_rate": 4.995814788271582e-05,
178
+ "loss": 2.0301,
179
+ "mean_token_accuracy": 0.697366189956665,
180
+ "num_tokens": 49807360.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.3566651805617477,
185
+ "grad_norm": 0.6190801858901978,
186
+ "learning_rate": 4.993932989141088e-05,
187
+ "loss": 1.9408,
188
+ "mean_token_accuracy": 0.7095360569655895,
189
+ "num_tokens": 52428800.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.37449843958983503,
194
+ "grad_norm": 0.6333149671554565,
195
+ "learning_rate": 4.991703320630011e-05,
196
+ "loss": 1.9076,
197
+ "mean_token_accuracy": 0.7127362459897995,
198
+ "num_tokens": 55050240.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.39233169861792244,
203
+ "grad_norm": 0.9046627879142761,
204
+ "learning_rate": 4.989126093746468e-05,
205
+ "loss": 1.8913,
206
+ "mean_token_accuracy": 0.7148448958992958,
207
+ "num_tokens": 57671680.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.4101649576460098,
212
+ "grad_norm": 0.7778757214546204,
213
+ "learning_rate": 4.986201667978193e-05,
214
+ "loss": 1.9016,
215
+ "mean_token_accuracy": 0.7118340134620667,
216
+ "num_tokens": 60293120.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.4279982166740972,
221
+ "grad_norm": 0.6346383690834045,
222
+ "learning_rate": 4.9829304512423944e-05,
223
+ "loss": 1.898,
224
+ "mean_token_accuracy": 0.7123855009675026,
225
+ "num_tokens": 62914560.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.44583147570218457,
230
+ "grad_norm": 0.7308981418609619,
231
+ "learning_rate": 4.979312899828853e-05,
232
+ "loss": 1.8435,
233
+ "mean_token_accuracy": 0.7184088006615639,
234
+ "num_tokens": 65536000.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.463664734730272,
239
+ "grad_norm": 0.6174861788749695,
240
+ "learning_rate": 4.9753495183362796e-05,
241
+ "loss": 1.8242,
242
+ "mean_token_accuracy": 0.7238672457635402,
243
+ "num_tokens": 68157440.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.48149799375835933,
248
+ "grad_norm": 0.6507575511932373,
249
+ "learning_rate": 4.971040859601927e-05,
250
+ "loss": 1.791,
251
+ "mean_token_accuracy": 0.7251060992479325,
252
+ "num_tokens": 70778880.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.49933125278644674,
257
+ "grad_norm": 0.8092382550239563,
258
+ "learning_rate": 4.966387524624482e-05,
259
+ "loss": 1.7879,
260
+ "mean_token_accuracy": 0.7255102708935738,
261
+ "num_tokens": 73400320.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.5171645118145342,
266
+ "grad_norm": 0.5636172294616699,
267
+ "learning_rate": 4.9613901624802245e-05,
268
+ "loss": 1.7351,
269
+ "mean_token_accuracy": 0.7330663032829762,
270
+ "num_tokens": 76021760.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.5349977708426215,
275
+ "grad_norm": 0.655262291431427,
276
+ "learning_rate": 4.956049470232503e-05,
277
+ "loss": 1.7464,
278
+ "mean_token_accuracy": 0.7302443347871304,
279
+ "num_tokens": 78643200.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.5528310298707089,
284
+ "grad_norm": 0.9413332343101501,
285
+ "learning_rate": 4.9503661928344944e-05,
286
+ "loss": 1.7277,
287
+ "mean_token_accuracy": 0.730958030372858,
288
+ "num_tokens": 81264640.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.5706642888987963,
293
+ "grad_norm": 0.7439999580383301,
294
+ "learning_rate": 4.9443411230252956e-05,
295
+ "loss": 1.6947,
296
+ "mean_token_accuracy": 0.7346593342721462,
297
+ "num_tokens": 83886080.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.5884975479268837,
302
+ "grad_norm": 0.7665299773216248,
303
+ "learning_rate": 4.93797510121935e-05,
304
+ "loss": 1.7136,
305
+ "mean_token_accuracy": 0.7324949607253075,
306
+ "num_tokens": 86507520.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.606330806954971,
311
+ "grad_norm": 0.7366372346878052,
312
+ "learning_rate": 4.931269015389217e-05,
313
+ "loss": 1.676,
314
+ "mean_token_accuracy": 0.7371297933161258,
315
+ "num_tokens": 89128960.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.6241640659830584,
320
+ "grad_norm": 0.6571912169456482,
321
+ "learning_rate": 4.9242238009417175e-05,
322
+ "loss": 1.6252,
323
+ "mean_token_accuracy": 0.7434912383556366,
324
+ "num_tokens": 91750400.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.6419973250111458,
329
+ "grad_norm": 0.7496657967567444,
330
+ "learning_rate": 4.9168404405874505e-05,
331
+ "loss": 1.657,
332
+ "mean_token_accuracy": 0.737925547361374,
333
+ "num_tokens": 94371840.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.6598305840392331,
338
+ "grad_norm": 0.5275835394859314,
339
+ "learning_rate": 4.9091199642037244e-05,
340
+ "loss": 1.6723,
341
+ "mean_token_accuracy": 0.7358783438801766,
342
+ "num_tokens": 96993280.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.6776638430673205,
347
+ "grad_norm": 0.6143308281898499,
348
+ "learning_rate": 4.901063448690901e-05,
349
+ "loss": 1.6292,
350
+ "mean_token_accuracy": 0.7412570238113403,
351
+ "num_tokens": 99614720.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.6954971020954079,
356
+ "grad_norm": 0.860007643699646,
357
+ "learning_rate": 4.892672017822183e-05,
358
+ "loss": 1.6281,
359
+ "mean_token_accuracy": 0.7422710858285427,
360
+ "num_tokens": 102236160.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.7133303611234953,
365
+ "grad_norm": 0.5295814275741577,
366
+ "learning_rate": 4.8839468420868606e-05,
367
+ "loss": 1.6112,
368
+ "mean_token_accuracy": 0.7446457460522652,
369
+ "num_tokens": 104857600.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.7311636201515826,
374
+ "grad_norm": 0.7366165518760681,
375
+ "learning_rate": 4.8748891385270495e-05,
376
+ "loss": 1.6197,
377
+ "mean_token_accuracy": 0.7417318046092987,
378
+ "num_tokens": 107479040.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.7489968791796701,
383
+ "grad_norm": 0.7033573985099792,
384
+ "learning_rate": 4.865500170567926e-05,
385
+ "loss": 1.5763,
386
+ "mean_token_accuracy": 0.7474127545952797,
387
+ "num_tokens": 110100480.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.7668301382077575,
392
+ "grad_norm": 0.6527330875396729,
393
+ "learning_rate": 4.855781247841498e-05,
394
+ "loss": 1.5714,
395
+ "mean_token_accuracy": 0.7477886863052845,
396
+ "num_tokens": 112721920.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.7846633972358449,
401
+ "grad_norm": 0.7344551682472229,
402
+ "learning_rate": 4.845733726003928e-05,
403
+ "loss": 1.5682,
404
+ "mean_token_accuracy": 0.7475516766309738,
405
+ "num_tokens": 115343360.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.8024966562639322,
410
+ "grad_norm": 0.6840972304344177,
411
+ "learning_rate": 4.835359006546443e-05,
412
+ "loss": 1.5955,
413
+ "mean_token_accuracy": 0.7427977688610554,
414
+ "num_tokens": 117964800.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.8203299152920196,
419
+ "grad_norm": 0.635522723197937,
420
+ "learning_rate": 4.824658536599836e-05,
421
+ "loss": 1.5459,
422
+ "mean_token_accuracy": 0.7526196874678135,
423
+ "num_tokens": 120586240.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.838163174320107,
428
+ "grad_norm": 0.6874827146530151,
429
+ "learning_rate": 4.8136338087326216e-05,
430
+ "loss": 1.6108,
431
+ "mean_token_accuracy": 0.7413325920701027,
432
+ "num_tokens": 123207680.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.8559964333481944,
437
+ "grad_norm": 0.6571274995803833,
438
+ "learning_rate": 4.8022863607428356e-05,
439
+ "loss": 1.539,
440
+ "mean_token_accuracy": 0.7521994858980179,
441
+ "num_tokens": 125829120.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.8738296923762817,
446
+ "grad_norm": 0.8841336369514465,
447
+ "learning_rate": 4.790617775443537e-05,
448
+ "loss": 1.5486,
449
+ "mean_token_accuracy": 0.7492702730000019,
450
+ "num_tokens": 128450560.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.8916629514043691,
455
+ "grad_norm": 0.690990149974823,
456
+ "learning_rate": 4.7786296804420274e-05,
457
+ "loss": 1.5646,
458
+ "mean_token_accuracy": 0.7455468386411667,
459
+ "num_tokens": 131072000.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.9094962104324565,
464
+ "grad_norm": 0.7215222716331482,
465
+ "learning_rate": 4.766323747912818e-05,
466
+ "loss": 1.5085,
467
+ "mean_token_accuracy": 0.754824897646904,
468
+ "num_tokens": 133693440.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.927329469460544,
473
+ "grad_norm": 0.6024391651153564,
474
+ "learning_rate": 4.753701694364392e-05,
475
+ "loss": 1.4856,
476
+ "mean_token_accuracy": 0.7570274345576763,
477
+ "num_tokens": 136314880.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.9451627284886313,
482
+ "grad_norm": 0.6131455302238464,
483
+ "learning_rate": 4.740765280399767e-05,
484
+ "loss": 1.5433,
485
+ "mean_token_accuracy": 0.7495988786220551,
486
+ "num_tokens": 138936320.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.9629959875167187,
491
+ "grad_norm": 0.7875815629959106,
492
+ "learning_rate": 4.72751631047092e-05,
493
+ "loss": 1.4917,
494
+ "mean_token_accuracy": 0.7559954360127449,
495
+ "num_tokens": 141557760.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.9808292465448061,
500
+ "grad_norm": 0.7112794518470764,
501
+ "learning_rate": 4.71395663262709e-05,
502
+ "loss": 1.4784,
503
+ "mean_token_accuracy": 0.7572461232542992,
504
+ "num_tokens": 144179200.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.9986625055728935,
509
+ "grad_norm": 0.6807007789611816,
510
+ "learning_rate": 4.7000881382570014e-05,
511
+ "loss": 1.5159,
512
+ "mean_token_accuracy": 0.7523536741733551,
513
+ "num_tokens": 146800640.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 1.0,
518
+ "eval_loss": 1.494895339012146,
519
+ "eval_mean_token_accuracy": 0.7405715703964233,
520
+ "eval_num_tokens": 146980864.0,
521
+ "eval_runtime": 2.6056,
522
+ "eval_samples_per_second": 27.633,
523
+ "eval_steps_per_second": 1.919,
524
+ "step": 561
525
+ },
526
+ {
527
+ "epoch": 1.0160499331252786,
528
+ "grad_norm": 0.5480894446372986,
529
+ "learning_rate": 4.685912761825038e-05,
530
+ "loss": 1.4428,
531
+ "mean_token_accuracy": 0.7618689170250525,
532
+ "num_tokens": 149340160.0,
533
+ "step": 570
534
+ },
535
+ {
536
+ "epoch": 1.033883192153366,
537
+ "grad_norm": 0.6192435622215271,
538
+ "learning_rate": 4.671432480601416e-05,
539
+ "loss": 1.4711,
540
+ "mean_token_accuracy": 0.7556988887488842,
541
+ "num_tokens": 151961600.0,
542
+ "step": 580
543
+ },
544
+ {
545
+ "epoch": 1.0517164511814534,
546
+ "grad_norm": 0.5988848209381104,
547
+ "learning_rate": 4.656649314386378e-05,
548
+ "loss": 1.433,
549
+ "mean_token_accuracy": 0.7611474089324475,
550
+ "num_tokens": 154583040.0,
551
+ "step": 590
552
+ },
553
+ {
554
+ "epoch": 1.0695497102095408,
555
+ "grad_norm": 0.7051531672477722,
556
+ "learning_rate": 4.641565325228464e-05,
557
+ "loss": 1.415,
558
+ "mean_token_accuracy": 0.7643376767635346,
559
+ "num_tokens": 157204480.0,
560
+ "step": 600
561
+ },
562
+ {
563
+ "epoch": 1.0873829692376282,
564
+ "grad_norm": 0.6170579791069031,
565
+ "learning_rate": 4.6261826171368774e-05,
566
+ "loss": 1.4674,
567
+ "mean_token_accuracy": 0.7571297205984593,
568
+ "num_tokens": 159825920.0,
569
+ "step": 610
570
+ },
571
+ {
572
+ "epoch": 1.1052162282657156,
573
+ "grad_norm": 0.5691844820976257,
574
+ "learning_rate": 4.6105033357880104e-05,
575
+ "loss": 1.407,
576
+ "mean_token_accuracy": 0.7646078862249851,
577
+ "num_tokens": 162447360.0,
578
+ "step": 620
579
+ },
580
+ {
581
+ "epoch": 1.123049487293803,
582
+ "grad_norm": 0.5824171900749207,
583
+ "learning_rate": 4.594529668226146e-05,
584
+ "loss": 1.4185,
585
+ "mean_token_accuracy": 0.7614389970898628,
586
+ "num_tokens": 165068800.0,
587
+ "step": 630
588
+ },
589
+ {
590
+ "epoch": 1.1408827463218902,
591
+ "grad_norm": 0.5765758156776428,
592
+ "learning_rate": 4.578263842558402e-05,
593
+ "loss": 1.42,
594
+ "mean_token_accuracy": 0.7633033894002438,
595
+ "num_tokens": 167690240.0,
596
+ "step": 640
597
+ },
598
+ {
599
+ "epoch": 1.1587160053499777,
600
+ "grad_norm": 0.5261256694793701,
601
+ "learning_rate": 4.5617081276439355e-05,
602
+ "loss": 1.4384,
603
+ "mean_token_accuracy": 0.7604428693652153,
604
+ "num_tokens": 170311680.0,
605
+ "step": 650
606
+ },
607
+ {
608
+ "epoch": 1.176549264378065,
609
+ "grad_norm": 0.913208544254303,
610
+ "learning_rate": 4.544864832777469e-05,
611
+ "loss": 1.4775,
612
+ "mean_token_accuracy": 0.754877183586359,
613
+ "num_tokens": 172933120.0,
614
+ "step": 660
615
+ },
616
+ {
617
+ "epoch": 1.1943825234061525,
618
+ "grad_norm": 0.5799976587295532,
619
+ "learning_rate": 4.527736307367179e-05,
620
+ "loss": 1.4387,
621
+ "mean_token_accuracy": 0.7589173935353756,
622
+ "num_tokens": 175554560.0,
623
+ "step": 670
624
+ },
625
+ {
626
+ "epoch": 1.2122157824342399,
627
+ "grad_norm": 0.635047435760498,
628
+ "learning_rate": 4.510324940606979e-05,
629
+ "loss": 1.4439,
630
+ "mean_token_accuracy": 0.7597284145653248,
631
+ "num_tokens": 178176000.0,
632
+ "step": 680
633
+ },
634
+ {
635
+ "epoch": 1.2300490414623273,
636
+ "grad_norm": 0.5884873270988464,
637
+ "learning_rate": 4.4926331611432687e-05,
638
+ "loss": 1.3835,
639
+ "mean_token_accuracy": 0.7659787967801094,
640
+ "num_tokens": 180797440.0,
641
+ "step": 690
642
+ },
643
+ {
644
+ "epoch": 1.2478823004904147,
645
+ "grad_norm": 0.7232898473739624,
646
+ "learning_rate": 4.4746634367361634e-05,
647
+ "loss": 1.3629,
648
+ "mean_token_accuracy": 0.7691854700446129,
649
+ "num_tokens": 183418880.0,
650
+ "step": 700
651
+ },
652
+ {
653
+ "epoch": 1.265715559518502,
654
+ "grad_norm": 0.6569238305091858,
655
+ "learning_rate": 4.4564182739152805e-05,
656
+ "loss": 1.397,
657
+ "mean_token_accuracy": 0.7651830442249775,
658
+ "num_tokens": 186040320.0,
659
+ "step": 710
660
+ },
661
+ {
662
+ "epoch": 1.2835488185465893,
663
+ "grad_norm": 0.8630965352058411,
664
+ "learning_rate": 4.437900217630109e-05,
665
+ "loss": 1.4256,
666
+ "mean_token_accuracy": 0.7605882868170738,
667
+ "num_tokens": 188661760.0,
668
+ "step": 720
669
+ },
670
+ {
671
+ "epoch": 1.3013820775746767,
672
+ "grad_norm": 0.7600955367088318,
673
+ "learning_rate": 4.419111850895028e-05,
674
+ "loss": 1.3975,
675
+ "mean_token_accuracy": 0.7634797111153603,
676
+ "num_tokens": 191283200.0,
677
+ "step": 730
678
+ },
679
+ {
680
+ "epoch": 1.3192153366027641,
681
+ "grad_norm": 0.6687177419662476,
682
+ "learning_rate": 4.4000557944290086e-05,
683
+ "loss": 1.3616,
684
+ "mean_token_accuracy": 0.7703544825315476,
685
+ "num_tokens": 193904640.0,
686
+ "step": 740
687
+ },
688
+ {
689
+ "epoch": 1.3370485956308515,
690
+ "grad_norm": 0.719222903251648,
691
+ "learning_rate": 4.3807347062900624e-05,
692
+ "loss": 1.3788,
693
+ "mean_token_accuracy": 0.7648544363677502,
694
+ "num_tokens": 196526080.0,
695
+ "step": 750
696
+ },
697
+ {
698
+ "epoch": 1.354881854658939,
699
+ "grad_norm": 0.5986538529396057,
700
+ "learning_rate": 4.361151281504474e-05,
701
+ "loss": 1.364,
702
+ "mean_token_accuracy": 0.770479665696621,
703
+ "num_tokens": 199147520.0,
704
+ "step": 760
705
+ },
706
+ {
707
+ "epoch": 1.3727151136870264,
708
+ "grad_norm": 0.6472992897033691,
709
+ "learning_rate": 4.3413082516908893e-05,
710
+ "loss": 1.4045,
711
+ "mean_token_accuracy": 0.7644098080694676,
712
+ "num_tokens": 201768960.0,
713
+ "step": 770
714
+ },
715
+ {
716
+ "epoch": 1.3905483727151138,
717
+ "grad_norm": 0.44522207975387573,
718
+ "learning_rate": 4.321208384679285e-05,
719
+ "loss": 1.3314,
720
+ "mean_token_accuracy": 0.773588253557682,
721
+ "num_tokens": 204390400.0,
722
+ "step": 780
723
+ },
724
+ {
725
+ "epoch": 1.4083816317432012,
726
+ "grad_norm": 0.5019171833992004,
727
+ "learning_rate": 4.3008544841248986e-05,
728
+ "loss": 1.3591,
729
+ "mean_token_accuracy": 0.7695087313652038,
730
+ "num_tokens": 207011840.0,
731
+ "step": 790
732
+ },
733
+ {
734
+ "epoch": 1.4262148907712884,
735
+ "grad_norm": 0.7287567257881165,
736
+ "learning_rate": 4.280249389117157e-05,
737
+ "loss": 1.3773,
738
+ "mean_token_accuracy": 0.7675351880490779,
739
+ "num_tokens": 209633280.0,
740
+ "step": 800
741
+ },
742
+ {
743
+ "epoch": 1.4440481497993758,
744
+ "grad_norm": 1.0294618606567383,
745
+ "learning_rate": 4.2593959737836604e-05,
746
+ "loss": 1.3719,
747
+ "mean_token_accuracy": 0.7668005004525185,
748
+ "num_tokens": 212254720.0,
749
+ "step": 810
750
+ },
751
+ {
752
+ "epoch": 1.4618814088274632,
753
+ "grad_norm": 0.6181490421295166,
754
+ "learning_rate": 4.2382971468892806e-05,
755
+ "loss": 1.3168,
756
+ "mean_token_accuracy": 0.7758259050548076,
757
+ "num_tokens": 214876160.0,
758
+ "step": 820
759
+ },
760
+ {
761
+ "epoch": 1.4797146678555506,
762
+ "grad_norm": 0.5945810079574585,
763
+ "learning_rate": 4.216955851430432e-05,
764
+ "loss": 1.3536,
765
+ "mean_token_accuracy": 0.7710937492549419,
766
+ "num_tokens": 217497600.0,
767
+ "step": 830
768
+ },
769
+ {
770
+ "epoch": 1.497547926883638,
771
+ "grad_norm": 0.6803934574127197,
772
+ "learning_rate": 4.195375064224562e-05,
773
+ "loss": 1.324,
774
+ "mean_token_accuracy": 0.7753530308604241,
775
+ "num_tokens": 220119040.0,
776
+ "step": 840
777
+ },
778
+ {
779
+ "epoch": 1.5153811859117252,
780
+ "grad_norm": 0.5748502612113953,
781
+ "learning_rate": 4.1735577954949225e-05,
782
+ "loss": 1.3057,
783
+ "mean_token_accuracy": 0.7755480580031872,
784
+ "num_tokens": 222740480.0,
785
+ "step": 850
786
+ },
787
+ {
788
+ "epoch": 1.5332144449398126,
789
+ "grad_norm": 0.5675086379051208,
790
+ "learning_rate": 4.151507088450692e-05,
791
+ "loss": 1.302,
792
+ "mean_token_accuracy": 0.7765964694321156,
793
+ "num_tokens": 225361920.0,
794
+ "step": 860
795
+ },
796
+ {
797
+ "epoch": 1.5510477039679,
798
+ "grad_norm": 0.6266924142837524,
799
+ "learning_rate": 4.129226018862484e-05,
800
+ "loss": 1.353,
801
+ "mean_token_accuracy": 0.770804837346077,
802
+ "num_tokens": 227983360.0,
803
+ "step": 870
804
+ },
805
+ {
806
+ "epoch": 1.5688809629959874,
807
+ "grad_norm": 0.551527202129364,
808
+ "learning_rate": 4.106717694633321e-05,
809
+ "loss": 1.3477,
810
+ "mean_token_accuracy": 0.77206010222435,
811
+ "num_tokens": 230604800.0,
812
+ "step": 880
813
+ },
814
+ {
815
+ "epoch": 1.5867142220240749,
816
+ "grad_norm": 0.6628434658050537,
817
+ "learning_rate": 4.0839852553651265e-05,
818
+ "loss": 1.3244,
819
+ "mean_token_accuracy": 0.773486353456974,
820
+ "num_tokens": 233226240.0,
821
+ "step": 890
822
+ },
823
+ {
824
+ "epoch": 1.6045474810521623,
825
+ "grad_norm": 0.5956223011016846,
826
+ "learning_rate": 4.0610318719207906e-05,
827
+ "loss": 1.3298,
828
+ "mean_token_accuracy": 0.7733359761536122,
829
+ "num_tokens": 235847680.0,
830
+ "step": 900
831
+ },
832
+ {
833
+ "epoch": 1.6223807400802497,
834
+ "grad_norm": 0.6573621034622192,
835
+ "learning_rate": 4.037860745981881e-05,
836
+ "loss": 1.283,
837
+ "mean_token_accuracy": 0.7787142746150494,
838
+ "num_tokens": 238469120.0,
839
+ "step": 910
840
+ },
841
+ {
842
+ "epoch": 1.640213999108337,
843
+ "grad_norm": 0.607742965221405,
844
+ "learning_rate": 4.01447510960205e-05,
845
+ "loss": 1.2955,
846
+ "mean_token_accuracy": 0.7780937045812607,
847
+ "num_tokens": 241090560.0,
848
+ "step": 920
849
+ },
850
+ {
851
+ "epoch": 1.6580472581364245,
852
+ "grad_norm": 0.6278334856033325,
853
+ "learning_rate": 3.99087822475621e-05,
854
+ "loss": 1.3215,
855
+ "mean_token_accuracy": 0.7741382211446762,
856
+ "num_tokens": 243712000.0,
857
+ "step": 930
858
+ },
859
+ {
860
+ "epoch": 1.675880517164512,
861
+ "grad_norm": 0.5402811169624329,
862
+ "learning_rate": 3.96707338288553e-05,
863
+ "loss": 1.3113,
864
+ "mean_token_accuracy": 0.7751454092562199,
865
+ "num_tokens": 246333440.0,
866
+ "step": 940
867
+ },
868
+ {
869
+ "epoch": 1.6937137761925993,
870
+ "grad_norm": 0.6165571212768555,
871
+ "learning_rate": 3.9430639044383274e-05,
872
+ "loss": 1.3093,
873
+ "mean_token_accuracy": 0.7768628627061844,
874
+ "num_tokens": 248954880.0,
875
+ "step": 950
876
+ },
877
+ {
878
+ "epoch": 1.7115470352206867,
879
+ "grad_norm": 0.6086046695709229,
880
+ "learning_rate": 3.9188531384069096e-05,
881
+ "loss": 1.3125,
882
+ "mean_token_accuracy": 0.7762621335685254,
883
+ "num_tokens": 251576320.0,
884
+ "step": 960
885
+ },
886
+ {
887
+ "epoch": 1.729380294248774,
888
+ "grad_norm": 0.682686984539032,
889
+ "learning_rate": 3.8944444618604356e-05,
890
+ "loss": 1.2821,
891
+ "mean_token_accuracy": 0.7795726217329502,
892
+ "num_tokens": 254197760.0,
893
+ "step": 970
894
+ },
895
+ {
896
+ "epoch": 1.7472135532768613,
897
+ "grad_norm": 0.5615099668502808,
898
+ "learning_rate": 3.869841279473862e-05,
899
+ "loss": 1.3095,
900
+ "mean_token_accuracy": 0.7766918800771236,
901
+ "num_tokens": 256819200.0,
902
+ "step": 980
903
+ },
904
+ {
905
+ "epoch": 1.7650468123049488,
906
+ "grad_norm": 0.5341463088989258,
907
+ "learning_rate": 3.845047023053037e-05,
908
+ "loss": 1.2864,
909
+ "mean_token_accuracy": 0.779373399913311,
910
+ "num_tokens": 259440640.0,
911
+ "step": 990
912
+ },
913
+ {
914
+ "epoch": 1.7828800713330362,
915
+ "grad_norm": 0.5641157627105713,
916
+ "learning_rate": 3.82006515105601e-05,
917
+ "loss": 1.2857,
918
+ "mean_token_accuracy": 0.7798695519566536,
919
+ "num_tokens": 262062080.0,
920
+ "step": 1000
921
+ },
922
+ {
923
+ "epoch": 1.8007133303611234,
924
+ "grad_norm": 0.5196830630302429,
925
+ "learning_rate": 3.794899148110628e-05,
926
+ "loss": 1.2949,
927
+ "mean_token_accuracy": 0.7788909852504731,
928
+ "num_tokens": 264683520.0,
929
+ "step": 1010
930
+ },
931
+ {
932
+ "epoch": 1.8185465893892108,
933
+ "grad_norm": 0.4990008771419525,
934
+ "learning_rate": 3.7695525245284716e-05,
935
+ "loss": 1.254,
936
+ "mean_token_accuracy": 0.7839147254824639,
937
+ "num_tokens": 267304960.0,
938
+ "step": 1020
939
+ },
940
+ {
941
+ "epoch": 1.8363798484172982,
942
+ "grad_norm": 0.5643882751464844,
943
+ "learning_rate": 3.7440288158152187e-05,
944
+ "loss": 1.2446,
945
+ "mean_token_accuracy": 0.785726061463356,
946
+ "num_tokens": 269926400.0,
947
+ "step": 1030
948
+ },
949
+ {
950
+ "epoch": 1.8542131074453856,
951
+ "grad_norm": 0.8107313513755798,
952
+ "learning_rate": 3.7183315821774914e-05,
953
+ "loss": 1.2974,
954
+ "mean_token_accuracy": 0.7775464087724686,
955
+ "num_tokens": 272547840.0,
956
+ "step": 1040
957
+ },
958
+ {
959
+ "epoch": 1.872046366473473,
960
+ "grad_norm": 0.5176706314086914,
961
+ "learning_rate": 3.692464408026253e-05,
962
+ "loss": 1.2916,
963
+ "mean_token_accuracy": 0.7772151328623295,
964
+ "num_tokens": 275169280.0,
965
+ "step": 1050
966
+ },
967
+ {
968
+ "epoch": 1.8898796255015604,
969
+ "grad_norm": 0.5898165106773376,
970
+ "learning_rate": 3.666430901476833e-05,
971
+ "loss": 1.2716,
972
+ "mean_token_accuracy": 0.7806546173989772,
973
+ "num_tokens": 277790720.0,
974
+ "step": 1060
975
+ },
976
+ {
977
+ "epoch": 1.9077128845296478,
978
+ "grad_norm": 0.5368467569351196,
979
+ "learning_rate": 3.640234693845645e-05,
980
+ "loss": 1.2491,
981
+ "mean_token_accuracy": 0.7844746112823486,
982
+ "num_tokens": 280412160.0,
983
+ "step": 1070
984
+ },
985
+ {
986
+ "epoch": 1.9255461435577352,
987
+ "grad_norm": 0.5310569405555725,
988
+ "learning_rate": 3.613879439143667e-05,
989
+ "loss": 1.2825,
990
+ "mean_token_accuracy": 0.7795199528336525,
991
+ "num_tokens": 283033600.0,
992
+ "step": 1080
993
+ },
994
+ {
995
+ "epoch": 1.9433794025858226,
996
+ "grad_norm": 0.6114910840988159,
997
+ "learning_rate": 3.58736881356676e-05,
998
+ "loss": 1.2595,
999
+ "mean_token_accuracy": 0.7831220239400863,
1000
+ "num_tokens": 285655040.0,
1001
+ "step": 1090
1002
+ },
1003
+ {
1004
+ "epoch": 1.96121266161391,
1005
+ "grad_norm": 0.5786811113357544,
1006
+ "learning_rate": 3.5607065149828843e-05,
1007
+ "loss": 1.2694,
1008
+ "mean_token_accuracy": 0.7822808526456356,
1009
+ "num_tokens": 288276480.0,
1010
+ "step": 1100
1011
+ },
1012
+ {
1013
+ "epoch": 1.9790459206419975,
1014
+ "grad_norm": 0.6175279021263123,
1015
+ "learning_rate": 3.533896262416302e-05,
1016
+ "loss": 1.2712,
1017
+ "mean_token_accuracy": 0.7803981445729733,
1018
+ "num_tokens": 290897920.0,
1019
+ "step": 1110
1020
+ },
1021
+ {
1022
+ "epoch": 1.9968791796700847,
1023
+ "grad_norm": 0.568803608417511,
1024
+ "learning_rate": 3.506941795528821e-05,
1025
+ "loss": 1.2576,
1026
+ "mean_token_accuracy": 0.7821117818355561,
1027
+ "num_tokens": 293519360.0,
1028
+ "step": 1120
1029
+ },
1030
+ {
1031
+ "epoch": 2.0,
1032
+ "eval_loss": 1.2395988702774048,
1033
+ "eval_mean_token_accuracy": 0.7735649704933166,
1034
+ "eval_num_tokens": 293961728.0,
1035
+ "eval_runtime": 2.8804,
1036
+ "eval_samples_per_second": 24.997,
1037
+ "eval_steps_per_second": 1.736,
1038
+ "step": 1122
1039
+ },
1040
+ {
1041
+ "epoch": 2.0142666072224698,
1042
+ "grad_norm": 0.6652095913887024,
1043
+ "learning_rate": 3.479846874098167e-05,
1044
+ "loss": 1.2741,
1045
+ "mean_token_accuracy": 0.7794669828353784,
1046
+ "num_tokens": 296058880.0,
1047
+ "step": 1130
1048
+ },
1049
+ {
1050
+ "epoch": 2.032099866250557,
1051
+ "grad_norm": 0.7581707835197449,
1052
+ "learning_rate": 3.452615277493544e-05,
1053
+ "loss": 1.199,
1054
+ "mean_token_accuracy": 0.7900658771395683,
1055
+ "num_tokens": 298680320.0,
1056
+ "step": 1140
1057
+ },
1058
+ {
1059
+ "epoch": 2.0499331252786446,
1060
+ "grad_norm": 0.5383137464523315,
1061
+ "learning_rate": 3.425250804148464e-05,
1062
+ "loss": 1.2304,
1063
+ "mean_token_accuracy": 0.7861943542957306,
1064
+ "num_tokens": 301301760.0,
1065
+ "step": 1150
1066
+ },
1067
+ {
1068
+ "epoch": 2.067766384306732,
1069
+ "grad_norm": 0.5952672958374023,
1070
+ "learning_rate": 3.3977572710309166e-05,
1071
+ "loss": 1.2445,
1072
+ "mean_token_accuracy": 0.7836834371089936,
1073
+ "num_tokens": 303923200.0,
1074
+ "step": 1160
1075
+ },
1076
+ {
1077
+ "epoch": 2.0855996433348194,
1078
+ "grad_norm": 0.646511435508728,
1079
+ "learning_rate": 3.3701385131109616e-05,
1080
+ "loss": 1.2019,
1081
+ "mean_token_accuracy": 0.7892128698527813,
1082
+ "num_tokens": 306544640.0,
1083
+ "step": 1170
1084
+ },
1085
+ {
1086
+ "epoch": 2.103432902362907,
1087
+ "grad_norm": 0.514665961265564,
1088
+ "learning_rate": 3.342398382825791e-05,
1089
+ "loss": 1.2614,
1090
+ "mean_token_accuracy": 0.7812080189585686,
1091
+ "num_tokens": 309166080.0,
1092
+ "step": 1180
1093
+ },
1094
+ {
1095
+ "epoch": 2.121266161390994,
1096
+ "grad_norm": 0.5490918159484863,
1097
+ "learning_rate": 3.31454074954238e-05,
1098
+ "loss": 1.2108,
1099
+ "mean_token_accuracy": 0.7876022845506668,
1100
+ "num_tokens": 311787520.0,
1101
+ "step": 1190
1102
+ },
1103
+ {
1104
+ "epoch": 2.1390994204190816,
1105
+ "grad_norm": 0.6549789309501648,
1106
+ "learning_rate": 3.286569499017752e-05,
1107
+ "loss": 1.2024,
1108
+ "mean_token_accuracy": 0.7898712314665317,
1109
+ "num_tokens": 314408960.0,
1110
+ "step": 1200
1111
+ },
1112
+ {
1113
+ "epoch": 2.156932679447169,
1114
+ "grad_norm": 0.5743107199668884,
1115
+ "learning_rate": 3.2584885328569785e-05,
1116
+ "loss": 1.1835,
1117
+ "mean_token_accuracy": 0.7928275369107723,
1118
+ "num_tokens": 317030400.0,
1119
+ "step": 1210
1120
+ },
1121
+ {
1122
+ "epoch": 2.1747659384752565,
1123
+ "grad_norm": 0.6201761960983276,
1124
+ "learning_rate": 3.2303017679689525e-05,
1125
+ "loss": 1.1939,
1126
+ "mean_token_accuracy": 0.7908047571778297,
1127
+ "num_tokens": 319651840.0,
1128
+ "step": 1220
1129
+ },
1130
+ {
1131
+ "epoch": 2.192599197503344,
1132
+ "grad_norm": 0.550830602645874,
1133
+ "learning_rate": 3.2020131360200354e-05,
1134
+ "loss": 1.1864,
1135
+ "mean_token_accuracy": 0.7930233284831048,
1136
+ "num_tokens": 322273280.0,
1137
+ "step": 1230
1138
+ },
1139
+ {
1140
+ "epoch": 2.2104324565314313,
1141
+ "grad_norm": 0.5972590446472168,
1142
+ "learning_rate": 3.173626582885645e-05,
1143
+ "loss": 1.2177,
1144
+ "mean_token_accuracy": 0.7885610058903694,
1145
+ "num_tokens": 324894720.0,
1146
+ "step": 1240
1147
+ },
1148
+ {
1149
+ "epoch": 2.2282657155595187,
1150
+ "grad_norm": 0.5069220066070557,
1151
+ "learning_rate": 3.145146068099858e-05,
1152
+ "loss": 1.2279,
1153
+ "mean_token_accuracy": 0.7862870991230011,
1154
+ "num_tokens": 327516160.0,
1155
+ "step": 1250
1156
+ },
1157
+ {
1158
+ "epoch": 2.246098974587606,
1159
+ "grad_norm": 0.5100940465927124,
1160
+ "learning_rate": 3.116575564303109e-05,
1161
+ "loss": 1.1777,
1162
+ "mean_token_accuracy": 0.7929233327507973,
1163
+ "num_tokens": 330137600.0,
1164
+ "step": 1260
1165
+ },
1166
+ {
1167
+ "epoch": 2.2639322336156935,
1168
+ "grad_norm": 0.6175646781921387,
1169
+ "learning_rate": 3.087919056688067e-05,
1170
+ "loss": 1.173,
1171
+ "mean_token_accuracy": 0.7950430497527122,
1172
+ "num_tokens": 332759040.0,
1173
+ "step": 1270
1174
+ },
1175
+ {
1176
+ "epoch": 2.2817654926437805,
1177
+ "grad_norm": 0.5933928489685059,
1178
+ "learning_rate": 3.059180542443746e-05,
1179
+ "loss": 1.2057,
1180
+ "mean_token_accuracy": 0.7882316343486309,
1181
+ "num_tokens": 335380480.0,
1182
+ "step": 1280
1183
+ },
1184
+ {
1185
+ "epoch": 2.299598751671868,
1186
+ "grad_norm": 0.4882568120956421,
1187
+ "learning_rate": 3.0303640301979635e-05,
1188
+ "loss": 1.1982,
1189
+ "mean_token_accuracy": 0.7903318881988526,
1190
+ "num_tokens": 338001920.0,
1191
+ "step": 1290
1192
+ },
1193
+ {
1194
+ "epoch": 2.3174320106999553,
1195
+ "grad_norm": 0.5135630369186401,
1196
+ "learning_rate": 3.0014735394581823e-05,
1197
+ "loss": 1.2045,
1198
+ "mean_token_accuracy": 0.7893250770866871,
1199
+ "num_tokens": 340623360.0,
1200
+ "step": 1300
1201
+ },
1202
+ {
1203
+ "epoch": 2.3352652697280427,
1204
+ "grad_norm": 0.5496548414230347,
1205
+ "learning_rate": 2.972513100050851e-05,
1206
+ "loss": 1.1883,
1207
+ "mean_token_accuracy": 0.791755847632885,
1208
+ "num_tokens": 343244800.0,
1209
+ "step": 1310
1210
+ },
1211
+ {
1212
+ "epoch": 2.35309852875613,
1213
+ "grad_norm": 0.5601444840431213,
1214
+ "learning_rate": 2.943486751559295e-05,
1215
+ "loss": 1.1263,
1216
+ "mean_token_accuracy": 0.7998171858489513,
1217
+ "num_tokens": 345866240.0,
1218
+ "step": 1320
1219
+ },
1220
+ {
1221
+ "epoch": 2.3709317877842175,
1222
+ "grad_norm": 0.5243608355522156,
1223
+ "learning_rate": 2.9143985427602516e-05,
1224
+ "loss": 1.2274,
1225
+ "mean_token_accuracy": 0.7866500526666641,
1226
+ "num_tokens": 348487680.0,
1227
+ "step": 1330
1228
+ },
1229
+ {
1230
+ "epoch": 2.388765046812305,
1231
+ "grad_norm": 0.573567807674408,
1232
+ "learning_rate": 2.88525253105912e-05,
1233
+ "loss": 1.1977,
1234
+ "mean_token_accuracy": 0.7901960171759128,
1235
+ "num_tokens": 351109120.0,
1236
+ "step": 1340
1237
+ },
1238
+ {
1239
+ "epoch": 2.4065983058403924,
1240
+ "grad_norm": 0.4823901951313019,
1241
+ "learning_rate": 2.856052781924012e-05,
1242
+ "loss": 1.1358,
1243
+ "mean_token_accuracy": 0.7998057343065739,
1244
+ "num_tokens": 353730560.0,
1245
+ "step": 1350
1246
+ },
1247
+ {
1248
+ "epoch": 2.4244315648684798,
1249
+ "grad_norm": 0.515941858291626,
1250
+ "learning_rate": 2.8268033683186697e-05,
1251
+ "loss": 1.1881,
1252
+ "mean_token_accuracy": 0.7909879550337792,
1253
+ "num_tokens": 356352000.0,
1254
+ "step": 1360
1255
+ },
1256
+ {
1257
+ "epoch": 2.442264823896567,
1258
+ "grad_norm": 0.5508431196212769,
1259
+ "learning_rate": 2.7975083701343497e-05,
1260
+ "loss": 1.171,
1261
+ "mean_token_accuracy": 0.7945770494639873,
1262
+ "num_tokens": 358973440.0,
1263
+ "step": 1370
1264
+ },
1265
+ {
1266
+ "epoch": 2.4600980829246546,
1267
+ "grad_norm": 0.5473787188529968,
1268
+ "learning_rate": 2.7681718736207298e-05,
1269
+ "loss": 1.1309,
1270
+ "mean_token_accuracy": 0.7995454497635365,
1271
+ "num_tokens": 361594880.0,
1272
+ "step": 1380
1273
+ },
1274
+ {
1275
+ "epoch": 2.477931341952742,
1276
+ "grad_norm": 0.6680080890655518,
1277
+ "learning_rate": 2.738797970815933e-05,
1278
+ "loss": 1.1848,
1279
+ "mean_token_accuracy": 0.7917638584971428,
1280
+ "num_tokens": 364216320.0,
1281
+ "step": 1390
1282
+ },
1283
+ {
1284
+ "epoch": 2.4957646009808294,
1285
+ "grad_norm": 0.49958300590515137,
1286
+ "learning_rate": 2.7093907589757468e-05,
1287
+ "loss": 1.1922,
1288
+ "mean_token_accuracy": 0.7899307683110237,
1289
+ "num_tokens": 366837760.0,
1290
+ "step": 1400
1291
+ },
1292
+ {
1293
+ "epoch": 2.5135978600089164,
1294
+ "grad_norm": 0.4601416289806366,
1295
+ "learning_rate": 2.6799543400021132e-05,
1296
+ "loss": 1.1448,
1297
+ "mean_token_accuracy": 0.7981104202568531,
1298
+ "num_tokens": 369459200.0,
1299
+ "step": 1410
1300
+ },
1301
+ {
1302
+ "epoch": 2.531431119037004,
1303
+ "grad_norm": 0.5619862079620361,
1304
+ "learning_rate": 2.650492819870965e-05,
1305
+ "loss": 1.1371,
1306
+ "mean_token_accuracy": 0.7994878143072128,
1307
+ "num_tokens": 372080640.0,
1308
+ "step": 1420
1309
+ },
1310
+ {
1311
+ "epoch": 2.549264378065091,
1312
+ "grad_norm": 0.6030861139297485,
1313
+ "learning_rate": 2.621010308059504e-05,
1314
+ "loss": 1.1695,
1315
+ "mean_token_accuracy": 0.7944411806762218,
1316
+ "num_tokens": 374702080.0,
1317
+ "step": 1430
1318
+ },
1319
+ {
1320
+ "epoch": 2.5670976370931786,
1321
+ "grad_norm": 0.5237954258918762,
1322
+ "learning_rate": 2.5915109169729814e-05,
1323
+ "loss": 1.1372,
1324
+ "mean_token_accuracy": 0.7982218623161316,
1325
+ "num_tokens": 377323520.0,
1326
+ "step": 1440
1327
+ },
1328
+ {
1329
+ "epoch": 2.584930896121266,
1330
+ "grad_norm": 0.5144650936126709,
1331
+ "learning_rate": 2.5619987613710756e-05,
1332
+ "loss": 1.1896,
1333
+ "mean_token_accuracy": 0.7915455557405948,
1334
+ "num_tokens": 379944960.0,
1335
+ "step": 1450
1336
+ },
1337
+ {
1338
+ "epoch": 2.6027641551493534,
1339
+ "grad_norm": 0.5404797196388245,
1340
+ "learning_rate": 2.5324779577939394e-05,
1341
+ "loss": 1.1364,
1342
+ "mean_token_accuracy": 0.7987703062593937,
1343
+ "num_tokens": 382566400.0,
1344
+ "step": 1460
1345
+ },
1346
+ {
1347
+ "epoch": 2.620597414177441,
1348
+ "grad_norm": 0.4770708382129669,
1349
+ "learning_rate": 2.5029526239880015e-05,
1350
+ "loss": 1.1541,
1351
+ "mean_token_accuracy": 0.7973127588629723,
1352
+ "num_tokens": 385187840.0,
1353
+ "step": 1470
1354
+ },
1355
+ {
1356
+ "epoch": 2.6384306732055283,
1357
+ "grad_norm": 0.5579174757003784,
1358
+ "learning_rate": 2.473426878331591e-05,
1359
+ "loss": 1.1695,
1360
+ "mean_token_accuracy": 0.7945644512772561,
1361
+ "num_tokens": 387809280.0,
1362
+ "step": 1480
1363
+ },
1364
+ {
1365
+ "epoch": 2.6562639322336157,
1366
+ "grad_norm": 0.5287127494812012,
1367
+ "learning_rate": 2.443904839260488e-05,
1368
+ "loss": 1.1615,
1369
+ "mean_token_accuracy": 0.7949182473123073,
1370
+ "num_tokens": 390430720.0,
1371
+ "step": 1490
1372
+ },
1373
+ {
1374
+ "epoch": 2.674097191261703,
1375
+ "grad_norm": 0.5106226801872253,
1376
+ "learning_rate": 2.4143906246934537e-05,
1377
+ "loss": 1.1201,
1378
+ "mean_token_accuracy": 0.8007564425468445,
1379
+ "num_tokens": 393052160.0,
1380
+ "step": 1500
1381
+ },
1382
+ {
1383
+ "epoch": 2.6919304502897905,
1384
+ "grad_norm": 0.48907479643821716,
1385
+ "learning_rate": 2.384888351457838e-05,
1386
+ "loss": 1.1908,
1387
+ "mean_token_accuracy": 0.7911803111433983,
1388
+ "num_tokens": 395673600.0,
1389
+ "step": 1510
1390
+ },
1391
+ {
1392
+ "epoch": 2.709763709317878,
1393
+ "grad_norm": 0.5045855045318604,
1394
+ "learning_rate": 2.3554021347153403e-05,
1395
+ "loss": 1.1483,
1396
+ "mean_token_accuracy": 0.7968017213046551,
1397
+ "num_tokens": 398295040.0,
1398
+ "step": 1520
1399
+ },
1400
+ {
1401
+ "epoch": 2.7275969683459653,
1402
+ "grad_norm": 0.505662739276886,
1403
+ "learning_rate": 2.3259360873879992e-05,
1404
+ "loss": 1.1749,
1405
+ "mean_token_accuracy": 0.7930756144225597,
1406
+ "num_tokens": 400916480.0,
1407
+ "step": 1530
1408
+ },
1409
+ {
1410
+ "epoch": 2.7454302273740527,
1411
+ "grad_norm": 0.5283376574516296,
1412
+ "learning_rate": 2.2964943195844966e-05,
1413
+ "loss": 1.173,
1414
+ "mean_token_accuracy": 0.7921638391911984,
1415
+ "num_tokens": 403537920.0,
1416
+ "step": 1540
1417
+ },
1418
+ {
1419
+ "epoch": 2.76326348640214,
1420
+ "grad_norm": 0.5634790062904358,
1421
+ "learning_rate": 2.2670809380268548e-05,
1422
+ "loss": 1.1779,
1423
+ "mean_token_accuracy": 0.7936561174690724,
1424
+ "num_tokens": 406159360.0,
1425
+ "step": 1550
1426
+ },
1427
+ {
1428
+ "epoch": 2.7810967454302276,
1429
+ "grad_norm": 0.48228803277015686,
1430
+ "learning_rate": 2.237700045477607e-05,
1431
+ "loss": 1.14,
1432
+ "mean_token_accuracy": 0.796902097016573,
1433
+ "num_tokens": 408780800.0,
1434
+ "step": 1560
1435
+ },
1436
+ {
1437
+ "epoch": 2.798930004458315,
1438
+ "grad_norm": 0.4753395617008209,
1439
+ "learning_rate": 2.2083557401675115e-05,
1440
+ "loss": 1.1294,
1441
+ "mean_token_accuracy": 0.7987142026424408,
1442
+ "num_tokens": 411402240.0,
1443
+ "step": 1570
1444
+ },
1445
+ {
1446
+ "epoch": 2.8167632634864024,
1447
+ "grad_norm": 0.5348941683769226,
1448
+ "learning_rate": 2.1790521152239137e-05,
1449
+ "loss": 1.1581,
1450
+ "mean_token_accuracy": 0.7944865994155407,
1451
+ "num_tokens": 414023680.0,
1452
+ "step": 1580
1453
+ },
1454
+ {
1455
+ "epoch": 2.83459652251449,
1456
+ "grad_norm": 0.5610408782958984,
1457
+ "learning_rate": 2.1497932580998053e-05,
1458
+ "loss": 1.1609,
1459
+ "mean_token_accuracy": 0.795001070201397,
1460
+ "num_tokens": 416645120.0,
1461
+ "step": 1590
1462
+ },
1463
+ {
1464
+ "epoch": 2.8524297815425768,
1465
+ "grad_norm": 0.5431333184242249,
1466
+ "learning_rate": 2.1205832500036838e-05,
1467
+ "loss": 1.1405,
1468
+ "mean_token_accuracy": 0.7983539171516896,
1469
+ "num_tokens": 419266560.0,
1470
+ "step": 1600
1471
+ },
1472
+ {
1473
+ "epoch": 2.870263040570664,
1474
+ "grad_norm": 0.5650882720947266,
1475
+ "learning_rate": 2.0914261653302747e-05,
1476
+ "loss": 1.17,
1477
+ "mean_token_accuracy": 0.7925355695188046,
1478
+ "num_tokens": 421888000.0,
1479
+ "step": 1610
1480
+ },
1481
+ {
1482
+ "epoch": 2.8880962995987516,
1483
+ "grad_norm": 0.5435991287231445,
1484
+ "learning_rate": 2.062326071092218e-05,
1485
+ "loss": 1.1562,
1486
+ "mean_token_accuracy": 0.7965150982141495,
1487
+ "num_tokens": 424509440.0,
1488
+ "step": 1620
1489
+ },
1490
+ {
1491
+ "epoch": 2.905929558626839,
1492
+ "grad_norm": 0.4466145932674408,
1493
+ "learning_rate": 2.0332870263527675e-05,
1494
+ "loss": 1.1199,
1495
+ "mean_token_accuracy": 0.8016460791230202,
1496
+ "num_tokens": 427130880.0,
1497
+ "step": 1630
1498
+ },
1499
+ {
1500
+ "epoch": 2.9237628176549264,
1501
+ "grad_norm": 0.5083450675010681,
1502
+ "learning_rate": 2.0043130816596136e-05,
1503
+ "loss": 1.1467,
1504
+ "mean_token_accuracy": 0.7959990955889225,
1505
+ "num_tokens": 429752320.0,
1506
+ "step": 1640
1507
+ },
1508
+ {
1509
+ "epoch": 2.941596076683014,
1510
+ "grad_norm": 0.49085912108421326,
1511
+ "learning_rate": 1.9754082784798843e-05,
1512
+ "loss": 1.1565,
1513
+ "mean_token_accuracy": 0.7959273479878902,
1514
+ "num_tokens": 432373760.0,
1515
+ "step": 1650
1516
+ },
1517
+ {
1518
+ "epoch": 2.9594293357111012,
1519
+ "grad_norm": 0.48444733023643494,
1520
+ "learning_rate": 1.9465766486364143e-05,
1521
+ "loss": 1.1342,
1522
+ "mean_token_accuracy": 0.797958905249834,
1523
+ "num_tokens": 434995200.0,
1524
+ "step": 1660
1525
+ },
1526
+ {
1527
+ "epoch": 2.9772625947391886,
1528
+ "grad_norm": 0.5343450903892517,
1529
+ "learning_rate": 1.9178222137453656e-05,
1530
+ "loss": 1.141,
1531
+ "mean_token_accuracy": 0.7985142156481743,
1532
+ "num_tokens": 437616640.0,
1533
+ "step": 1670
1534
+ },
1535
+ {
1536
+ "epoch": 2.995095853767276,
1537
+ "grad_norm": 0.5583868026733398,
1538
+ "learning_rate": 1.8891489846552646e-05,
1539
+ "loss": 1.1432,
1540
+ "mean_token_accuracy": 0.7971971161663532,
1541
+ "num_tokens": 440238080.0,
1542
+ "step": 1680
1543
+ },
1544
+ {
1545
+ "epoch": 3.0,
1546
+ "eval_loss": 1.1234296560287476,
1547
+ "eval_mean_token_accuracy": 0.7894601821899414,
1548
+ "eval_num_tokens": 440942592.0,
1549
+ "eval_runtime": 2.9179,
1550
+ "eval_samples_per_second": 24.675,
1551
+ "eval_steps_per_second": 1.714,
1552
+ "step": 1683
1553
+ },
1554
+ {
1555
+ "epoch": 3.0124832813196614,
1556
+ "grad_norm": 0.5347409844398499,
1557
+ "learning_rate": 1.860560960887546e-05,
1558
+ "loss": 1.1253,
1559
+ "mean_token_accuracy": 0.8009015726737487,
1560
+ "num_tokens": 442777600.0,
1561
+ "step": 1690
1562
+ },
1563
+ {
1564
+ "epoch": 3.0303165403477483,
1565
+ "grad_norm": 0.5598998069763184,
1566
+ "learning_rate": 1.8320621300786688e-05,
1567
+ "loss": 1.106,
1568
+ "mean_token_accuracy": 0.8023651242256165,
1569
+ "num_tokens": 445399040.0,
1570
+ "step": 1700
1571
+ },
1572
+ {
1573
+ "epoch": 3.0481497993758357,
1574
+ "grad_norm": 0.5113139152526855,
1575
+ "learning_rate": 1.8036564674239043e-05,
1576
+ "loss": 1.0887,
1577
+ "mean_token_accuracy": 0.8040673099458218,
1578
+ "num_tokens": 448020480.0,
1579
+ "step": 1710
1580
+ },
1581
+ {
1582
+ "epoch": 3.065983058403923,
1583
+ "grad_norm": 0.44441869854927063,
1584
+ "learning_rate": 1.7753479351228444e-05,
1585
+ "loss": 1.1185,
1586
+ "mean_token_accuracy": 0.7999229028820991,
1587
+ "num_tokens": 450641920.0,
1588
+ "step": 1720
1589
+ },
1590
+ {
1591
+ "epoch": 3.0838163174320106,
1592
+ "grad_norm": 0.47339409589767456,
1593
+ "learning_rate": 1.7471404818267316e-05,
1594
+ "loss": 1.0925,
1595
+ "mean_token_accuracy": 0.8034070417284965,
1596
+ "num_tokens": 453263360.0,
1597
+ "step": 1730
1598
+ },
1599
+ {
1600
+ "epoch": 3.101649576460098,
1601
+ "grad_norm": 0.47856712341308594,
1602
+ "learning_rate": 1.7190380420876766e-05,
1603
+ "loss": 1.0911,
1604
+ "mean_token_accuracy": 0.8045188069343567,
1605
+ "num_tokens": 455884800.0,
1606
+ "step": 1740
1607
+ },
1608
+ {
1609
+ "epoch": 3.1194828354881854,
1610
+ "grad_norm": 0.4918275773525238,
1611
+ "learning_rate": 1.6910445358098415e-05,
1612
+ "loss": 1.1109,
1613
+ "mean_token_accuracy": 0.8015380732715129,
1614
+ "num_tokens": 458506240.0,
1615
+ "step": 1750
1616
+ },
1617
+ {
1618
+ "epoch": 3.137316094516273,
1619
+ "grad_norm": 0.5043804049491882,
1620
+ "learning_rate": 1.6631638677026674e-05,
1621
+ "loss": 1.1364,
1622
+ "mean_token_accuracy": 0.7974261090159416,
1623
+ "num_tokens": 461127680.0,
1624
+ "step": 1760
1625
+ },
1626
+ {
1627
+ "epoch": 3.15514935354436,
1628
+ "grad_norm": 0.4904772937297821,
1629
+ "learning_rate": 1.6353999267362237e-05,
1630
+ "loss": 1.0838,
1631
+ "mean_token_accuracy": 0.806386636942625,
1632
+ "num_tokens": 463749120.0,
1633
+ "step": 1770
1634
+ },
1635
+ {
1636
+ "epoch": 3.1729826125724476,
1637
+ "grad_norm": 0.48294153809547424,
1638
+ "learning_rate": 1.607756585598743e-05,
1639
+ "loss": 1.1389,
1640
+ "mean_token_accuracy": 0.7980314172804356,
1641
+ "num_tokens": 466370560.0,
1642
+ "step": 1780
1643
+ },
1644
+ {
1645
+ "epoch": 3.190815871600535,
1646
+ "grad_norm": 0.5139598250389099,
1647
+ "learning_rate": 1.580237700156443e-05,
1648
+ "loss": 1.1095,
1649
+ "mean_token_accuracy": 0.8025780878961086,
1650
+ "num_tokens": 468992000.0,
1651
+ "step": 1790
1652
+ },
1653
+ {
1654
+ "epoch": 3.2086491306286224,
1655
+ "grad_norm": 0.47298020124435425,
1656
+ "learning_rate": 1.5528471089156804e-05,
1657
+ "loss": 1.0931,
1658
+ "mean_token_accuracy": 0.8041585251688957,
1659
+ "num_tokens": 471613440.0,
1660
+ "step": 1800
1661
+ },
1662
+ {
1663
+ "epoch": 3.22648238965671,
1664
+ "grad_norm": 0.41903576254844666,
1665
+ "learning_rate": 1.5255886324875363e-05,
1666
+ "loss": 1.0954,
1667
+ "mean_token_accuracy": 0.8047371134161949,
1668
+ "num_tokens": 474234880.0,
1669
+ "step": 1810
1670
+ },
1671
+ {
1672
+ "epoch": 3.2443156486847973,
1673
+ "grad_norm": 0.4610045552253723,
1674
+ "learning_rate": 1.4984660730548856e-05,
1675
+ "loss": 1.1095,
1676
+ "mean_token_accuracy": 0.8022483348846435,
1677
+ "num_tokens": 476856320.0,
1678
+ "step": 1820
1679
+ },
1680
+ {
1681
+ "epoch": 3.2621489077128847,
1682
+ "grad_norm": 0.5105413198471069,
1683
+ "learning_rate": 1.4714832138420553e-05,
1684
+ "loss": 1.1258,
1685
+ "mean_token_accuracy": 0.7983997173607349,
1686
+ "num_tokens": 479477760.0,
1687
+ "step": 1830
1688
+ },
1689
+ {
1690
+ "epoch": 3.279982166740972,
1691
+ "grad_norm": 0.4393883943557739,
1692
+ "learning_rate": 1.4446438185871092e-05,
1693
+ "loss": 1.0566,
1694
+ "mean_token_accuracy": 0.8093036323785782,
1695
+ "num_tokens": 482099200.0,
1696
+ "step": 1840
1697
+ },
1698
+ {
1699
+ "epoch": 3.297815425769059,
1700
+ "grad_norm": 0.4618268311023712,
1701
+ "learning_rate": 1.4179516310168623e-05,
1702
+ "loss": 1.0697,
1703
+ "mean_token_accuracy": 0.8082189694046974,
1704
+ "num_tokens": 484720640.0,
1705
+ "step": 1850
1706
+ },
1707
+ {
1708
+ "epoch": 3.3156486847971465,
1709
+ "grad_norm": 0.47192856669425964,
1710
+ "learning_rate": 1.3914103743246757e-05,
1711
+ "loss": 1.1203,
1712
+ "mean_token_accuracy": 0.8002961635589599,
1713
+ "num_tokens": 487342080.0,
1714
+ "step": 1860
1715
+ },
1716
+ {
1717
+ "epoch": 3.333481943825234,
1718
+ "grad_norm": 0.4617849290370941,
1719
+ "learning_rate": 1.3650237506511331e-05,
1720
+ "loss": 1.0664,
1721
+ "mean_token_accuracy": 0.8089914366602897,
1722
+ "num_tokens": 489963520.0,
1723
+ "step": 1870
1724
+ },
1725
+ {
1726
+ "epoch": 3.3513152028533213,
1727
+ "grad_norm": 0.46263769268989563,
1728
+ "learning_rate": 1.3387954405676329e-05,
1729
+ "loss": 1.0855,
1730
+ "mean_token_accuracy": 0.8045512497425079,
1731
+ "num_tokens": 492584960.0,
1732
+ "step": 1880
1733
+ },
1734
+ {
1735
+ "epoch": 3.3691484618814087,
1736
+ "grad_norm": 0.47842440009117126,
1737
+ "learning_rate": 1.3127291025630073e-05,
1738
+ "loss": 1.0705,
1739
+ "mean_token_accuracy": 0.8085929870605468,
1740
+ "num_tokens": 495206400.0,
1741
+ "step": 1890
1742
+ },
1743
+ {
1744
+ "epoch": 3.386981720909496,
1745
+ "grad_norm": 0.43500861525535583,
1746
+ "learning_rate": 1.2868283725332028e-05,
1747
+ "loss": 1.0635,
1748
+ "mean_token_accuracy": 0.8091395162045956,
1749
+ "num_tokens": 497827840.0,
1750
+ "step": 1900
1751
+ },
1752
+ {
1753
+ "epoch": 3.4048149799375835,
1754
+ "grad_norm": 0.508797287940979,
1755
+ "learning_rate": 1.261096863274136e-05,
1756
+ "loss": 1.1052,
1757
+ "mean_token_accuracy": 0.8021300226449967,
1758
+ "num_tokens": 500449280.0,
1759
+ "step": 1910
1760
+ },
1761
+ {
1762
+ "epoch": 3.422648238965671,
1763
+ "grad_norm": 0.42378219962120056,
1764
+ "learning_rate": 1.2355381639777472e-05,
1765
+ "loss": 1.0936,
1766
+ "mean_token_accuracy": 0.8043497301638126,
1767
+ "num_tokens": 503070720.0,
1768
+ "step": 1920
1769
+ },
1770
+ {
1771
+ "epoch": 3.4404814979937584,
1772
+ "grad_norm": 0.4547331631183624,
1773
+ "learning_rate": 1.2101558397313634e-05,
1774
+ "loss": 1.1142,
1775
+ "mean_token_accuracy": 0.8015869230031967,
1776
+ "num_tokens": 505692160.0,
1777
+ "step": 1930
1778
+ },
1779
+ {
1780
+ "epoch": 3.4583147570218458,
1781
+ "grad_norm": 0.4319451153278351,
1782
+ "learning_rate": 1.1849534310204152e-05,
1783
+ "loss": 1.0835,
1784
+ "mean_token_accuracy": 0.8058400988578797,
1785
+ "num_tokens": 508313600.0,
1786
+ "step": 1940
1787
+ },
1788
+ {
1789
+ "epoch": 3.476148016049933,
1790
+ "grad_norm": 0.4666835069656372,
1791
+ "learning_rate": 1.1599344532345877e-05,
1792
+ "loss": 1.1072,
1793
+ "mean_token_accuracy": 0.801591893285513,
1794
+ "num_tokens": 510935040.0,
1795
+ "step": 1950
1796
+ },
1797
+ {
1798
+ "epoch": 3.4939812750780206,
1799
+ "grad_norm": 0.47252121567726135,
1800
+ "learning_rate": 1.1351023961774749e-05,
1801
+ "loss": 1.1245,
1802
+ "mean_token_accuracy": 0.7985180281102657,
1803
+ "num_tokens": 513556480.0,
1804
+ "step": 1960
1805
+ },
1806
+ {
1807
+ "epoch": 3.511814534106108,
1808
+ "grad_norm": 0.5152117013931274,
1809
+ "learning_rate": 1.110460723579796e-05,
1810
+ "loss": 1.0982,
1811
+ "mean_token_accuracy": 0.8033757455646992,
1812
+ "num_tokens": 516177920.0,
1813
+ "step": 1970
1814
+ },
1815
+ {
1816
+ "epoch": 3.5296477931341954,
1817
+ "grad_norm": 0.43357202410697937,
1818
+ "learning_rate": 1.0860128726162545e-05,
1819
+ "loss": 1.0632,
1820
+ "mean_token_accuracy": 0.8095081992447376,
1821
+ "num_tokens": 518799360.0,
1822
+ "step": 1980
1823
+ },
1824
+ {
1825
+ "epoch": 3.547481052162283,
1826
+ "grad_norm": 0.45368409156799316,
1827
+ "learning_rate": 1.0617622534260985e-05,
1828
+ "loss": 1.114,
1829
+ "mean_token_accuracy": 0.8014212884008884,
1830
+ "num_tokens": 521420800.0,
1831
+ "step": 1990
1832
+ },
1833
+ {
1834
+ "epoch": 3.5653143111903702,
1835
+ "grad_norm": 0.38939619064331055,
1836
+ "learning_rate": 1.0377122486374576e-05,
1837
+ "loss": 1.0449,
1838
+ "mean_token_accuracy": 0.8123023062944412,
1839
+ "num_tokens": 524042240.0,
1840
+ "step": 2000
1841
+ },
1842
+ {
1843
+ "epoch": 3.5831475702184576,
1844
+ "grad_norm": 0.42768579721450806,
1845
+ "learning_rate": 1.0138662128955053e-05,
1846
+ "loss": 1.104,
1847
+ "mean_token_accuracy": 0.8025334358215332,
1848
+ "num_tokens": 526663680.0,
1849
+ "step": 2010
1850
+ },
1851
+ {
1852
+ "epoch": 3.600980829246545,
1853
+ "grad_norm": 0.41337716579437256,
1854
+ "learning_rate": 9.902274723945381e-06,
1855
+ "loss": 1.0788,
1856
+ "mean_token_accuracy": 0.8076182380318642,
1857
+ "num_tokens": 529285120.0,
1858
+ "step": 2020
1859
+ },
1860
+ {
1861
+ "epoch": 3.618814088274632,
1862
+ "grad_norm": 0.4830559194087982,
1863
+ "learning_rate": 9.667993244140134e-06,
1864
+ "loss": 1.0967,
1865
+ "mean_token_accuracy": 0.8028597474098206,
1866
+ "num_tokens": 531906560.0,
1867
+ "step": 2030
1868
+ },
1869
+ {
1870
+ "epoch": 3.6366473473027194,
1871
+ "grad_norm": 0.4195765554904938,
1872
+ "learning_rate": 9.435850368586286e-06,
1873
+ "loss": 1.0773,
1874
+ "mean_token_accuracy": 0.8066488340497017,
1875
+ "num_tokens": 534528000.0,
1876
+ "step": 2040
1877
+ },
1878
+ {
1879
+ "epoch": 3.654480606330807,
1880
+ "grad_norm": 0.4251314103603363,
1881
+ "learning_rate": 9.20587847802488e-06,
1882
+ "loss": 1.0819,
1883
+ "mean_token_accuracy": 0.8052599862217903,
1884
+ "num_tokens": 537149440.0,
1885
+ "step": 2050
1886
+ },
1887
+ {
1888
+ "epoch": 3.6723138653588943,
1889
+ "grad_norm": 0.4114632308483124,
1890
+ "learning_rate": 8.978109650374397e-06,
1891
+ "loss": 1.0697,
1892
+ "mean_token_accuracy": 0.8077415138483047,
1893
+ "num_tokens": 539770880.0,
1894
+ "step": 2060
1895
+ },
1896
+ {
1897
+ "epoch": 3.6901471243869817,
1898
+ "grad_norm": 0.4304318428039551,
1899
+ "learning_rate": 8.752575656256321e-06,
1900
+ "loss": 1.1046,
1901
+ "mean_token_accuracy": 0.8016697481274605,
1902
+ "num_tokens": 542392320.0,
1903
+ "step": 2070
1904
+ },
1905
+ {
1906
+ "epoch": 3.707980383415069,
1907
+ "grad_norm": 0.47687673568725586,
1908
+ "learning_rate": 8.52930795456355e-06,
1909
+ "loss": 1.0558,
1910
+ "mean_token_accuracy": 0.8107806429266929,
1911
+ "num_tokens": 545013760.0,
1912
+ "step": 2080
1913
+ },
1914
+ {
1915
+ "epoch": 3.7258136424431565,
1916
+ "grad_norm": 0.45524051785469055,
1917
+ "learning_rate": 8.30833768807237e-06,
1918
+ "loss": 1.0816,
1919
+ "mean_token_accuracy": 0.8059305518865585,
1920
+ "num_tokens": 547635200.0,
1921
+ "step": 2090
1922
+ },
1923
+ {
1924
+ "epoch": 3.743646901471244,
1925
+ "grad_norm": 0.45478421449661255,
1926
+ "learning_rate": 8.089695679098375e-06,
1927
+ "loss": 1.0961,
1928
+ "mean_token_accuracy": 0.8029544025659561,
1929
+ "num_tokens": 550256640.0,
1930
+ "step": 2100
1931
+ }
1932
+ ],
1933
+ "logging_steps": 10,
1934
+ "max_steps": 2800,
1935
+ "num_input_tokens_seen": 0,
1936
+ "num_train_epochs": 5,
1937
+ "save_steps": 140,
1938
+ "stateful_callbacks": {
1939
+ "TrainerControl": {
1940
+ "args": {
1941
+ "should_epoch_stop": false,
1942
+ "should_evaluate": false,
1943
+ "should_log": false,
1944
+ "should_save": true,
1945
+ "should_training_stop": false
1946
+ },
1947
+ "attributes": {}
1948
+ }
1949
+ },
1950
+ "total_flos": 2.7397788097786675e+18,
1951
+ "train_batch_size": 16,
1952
+ "trial_name": null,
1953
+ "trial_params": null
1954
+ }
checkpoint-2100/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7409b037df1da3374237680bbedeb80953c01e549dccfb71a127e04d8bc7415b
3
+ size 7352
checkpoint-2100/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2240/config.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_configs": [
3
+ {
4
+ "vocab_size": 49152,
5
+ "max_position_embeddings": 8192,
6
+ "hidden_size": 960,
7
+ "intermediate_size": 2560,
8
+ "num_hidden_layers": 32,
9
+ "num_attention_heads": 15,
10
+ "num_key_value_heads": 5,
11
+ "hidden_act": "silu",
12
+ "initializer_range": 0.02,
13
+ "rms_norm_eps": 1e-05,
14
+ "pretraining_tp": 1,
15
+ "use_cache": true,
16
+ "rope_theta": 100000,
17
+ "rope_scaling": null,
18
+ "attention_bias": false,
19
+ "attention_dropout": 0.0,
20
+ "mlp_bias": false,
21
+ "head_dim": 64,
22
+ "return_dict": true,
23
+ "output_hidden_states": false,
24
+ "output_attentions": false,
25
+ "torchscript": false,
26
+ "torch_dtype": "float32",
27
+ "use_bfloat16": false,
28
+ "tf_legacy_loss": false,
29
+ "pruned_heads": {},
30
+ "tie_word_embeddings": true,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "is_decoder": false,
34
+ "cross_attention_hidden_size": null,
35
+ "add_cross_attention": false,
36
+ "tie_encoder_decoder": false,
37
+ "max_length": 20,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "early_stopping": false,
41
+ "num_beams": 1,
42
+ "num_beam_groups": 1,
43
+ "diversity_penalty": 0.0,
44
+ "temperature": 1.0,
45
+ "top_k": 50,
46
+ "top_p": 1.0,
47
+ "typical_p": 1.0,
48
+ "repetition_penalty": 1.0,
49
+ "length_penalty": 1.0,
50
+ "no_repeat_ngram_size": 0,
51
+ "encoder_no_repeat_ngram_size": 0,
52
+ "bad_words_ids": null,
53
+ "num_return_sequences": 1,
54
+ "output_scores": false,
55
+ "return_dict_in_generate": false,
56
+ "forced_bos_token_id": null,
57
+ "forced_eos_token_id": null,
58
+ "remove_invalid_values": false,
59
+ "exponential_decay_length_penalty": null,
60
+ "suppress_tokens": null,
61
+ "begin_suppress_tokens": null,
62
+ "architectures": [
63
+ "LlamaForCausalLM"
64
+ ],
65
+ "finetuning_task": null,
66
+ "id2label": {
67
+ "0": "LABEL_0",
68
+ "1": "LABEL_1"
69
+ },
70
+ "label2id": {
71
+ "LABEL_0": 0,
72
+ "LABEL_1": 1
73
+ },
74
+ "tokenizer_class": null,
75
+ "prefix": null,
76
+ "bos_token_id": 1,
77
+ "pad_token_id": 2,
78
+ "eos_token_id": 2,
79
+ "sep_token_id": null,
80
+ "decoder_start_token_id": null,
81
+ "task_specific_params": null,
82
+ "problem_type": null,
83
+ "_name_or_path": "HuggingFaceTB/SmolLM2-360M-Instruct",
84
+ "_attn_implementation_autoset": true,
85
+ "transformers_version": "4.51.3",
86
+ "is_llama_config": true,
87
+ "model_type": "llama",
88
+ "rope_interleaved": false,
89
+ "transformers.js_config": {
90
+ "kv_cache_dtype": {
91
+ "q4f16": "float16",
92
+ "fp16": "float16"
93
+ }
94
+ }
95
+ },
96
+ {
97
+ "vocab_size": 49152,
98
+ "max_position_embeddings": 8192,
99
+ "hidden_size": 576,
100
+ "intermediate_size": 1536,
101
+ "num_hidden_layers": 30,
102
+ "num_attention_heads": 9,
103
+ "num_key_value_heads": 3,
104
+ "hidden_act": "silu",
105
+ "initializer_range": 0.041666666666666664,
106
+ "rms_norm_eps": 1e-05,
107
+ "pretraining_tp": 1,
108
+ "use_cache": true,
109
+ "rope_theta": 100000,
110
+ "rope_scaling": null,
111
+ "attention_bias": false,
112
+ "attention_dropout": 0.0,
113
+ "mlp_bias": false,
114
+ "head_dim": 64,
115
+ "return_dict": true,
116
+ "output_hidden_states": false,
117
+ "output_attentions": false,
118
+ "torchscript": false,
119
+ "torch_dtype": "float32",
120
+ "use_bfloat16": false,
121
+ "tf_legacy_loss": false,
122
+ "pruned_heads": {},
123
+ "tie_word_embeddings": true,
124
+ "chunk_size_feed_forward": 0,
125
+ "is_encoder_decoder": false,
126
+ "is_decoder": false,
127
+ "cross_attention_hidden_size": null,
128
+ "add_cross_attention": false,
129
+ "tie_encoder_decoder": false,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "do_sample": false,
133
+ "early_stopping": false,
134
+ "num_beams": 1,
135
+ "num_beam_groups": 1,
136
+ "diversity_penalty": 0.0,
137
+ "temperature": 1.0,
138
+ "top_k": 50,
139
+ "top_p": 1.0,
140
+ "typical_p": 1.0,
141
+ "repetition_penalty": 1.0,
142
+ "length_penalty": 1.0,
143
+ "no_repeat_ngram_size": 0,
144
+ "encoder_no_repeat_ngram_size": 0,
145
+ "bad_words_ids": null,
146
+ "num_return_sequences": 1,
147
+ "output_scores": false,
148
+ "return_dict_in_generate": false,
149
+ "forced_bos_token_id": null,
150
+ "forced_eos_token_id": null,
151
+ "remove_invalid_values": false,
152
+ "exponential_decay_length_penalty": null,
153
+ "suppress_tokens": null,
154
+ "begin_suppress_tokens": null,
155
+ "architectures": [
156
+ "LlamaForCausalLM"
157
+ ],
158
+ "finetuning_task": null,
159
+ "id2label": {
160
+ "0": "LABEL_0",
161
+ "1": "LABEL_1"
162
+ },
163
+ "label2id": {
164
+ "LABEL_0": 0,
165
+ "LABEL_1": 1
166
+ },
167
+ "tokenizer_class": null,
168
+ "prefix": null,
169
+ "bos_token_id": 1,
170
+ "pad_token_id": 2,
171
+ "eos_token_id": 2,
172
+ "sep_token_id": null,
173
+ "decoder_start_token_id": null,
174
+ "task_specific_params": null,
175
+ "problem_type": null,
176
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct",
177
+ "_attn_implementation_autoset": true,
178
+ "transformers_version": "4.51.3",
179
+ "is_llama_config": true,
180
+ "model_type": "llama",
181
+ "rope_interleaved": false,
182
+ "transformers.js_config": {
183
+ "kv_cache_dtype": {
184
+ "q4f16": "float16",
185
+ "fp16": "float16"
186
+ }
187
+ }
188
+ }
189
+ ],
190
+ "num_models": 2
191
+ }
checkpoint-2240/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2240/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7d0e428eaaebbd39500df7b70c85e0c423ad7178fcc03acaa80860c51ef011a
3
+ size 14244
checkpoint-2240/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df85220313a1a4e66213c606a0fb3019ca330ac0dcd88903b48308fd6db3b86a
3
+ size 1064
checkpoint-2240/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
checkpoint-2240/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2240/tokenizer_config.json ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<repo_name>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "4": {
37
+ "content": "<reponame>",
38
+ "lstrip": false,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "5": {
45
+ "content": "<file_sep>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "6": {
53
+ "content": "<filename>",
54
+ "lstrip": false,
55
+ "normalized": false,
56
+ "rstrip": false,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "7": {
61
+ "content": "<gh_stars>",
62
+ "lstrip": false,
63
+ "normalized": false,
64
+ "rstrip": false,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "8": {
69
+ "content": "<issue_start>",
70
+ "lstrip": false,
71
+ "normalized": false,
72
+ "rstrip": false,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "9": {
77
+ "content": "<issue_comment>",
78
+ "lstrip": false,
79
+ "normalized": false,
80
+ "rstrip": false,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "10": {
85
+ "content": "<issue_closed>",
86
+ "lstrip": false,
87
+ "normalized": false,
88
+ "rstrip": false,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "11": {
93
+ "content": "<jupyter_start>",
94
+ "lstrip": false,
95
+ "normalized": false,
96
+ "rstrip": false,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "12": {
101
+ "content": "<jupyter_text>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "13": {
109
+ "content": "<jupyter_code>",
110
+ "lstrip": false,
111
+ "normalized": false,
112
+ "rstrip": false,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "14": {
117
+ "content": "<jupyter_output>",
118
+ "lstrip": false,
119
+ "normalized": false,
120
+ "rstrip": false,
121
+ "single_word": false,
122
+ "special": true
123
+ },
124
+ "15": {
125
+ "content": "<jupyter_script>",
126
+ "lstrip": false,
127
+ "normalized": false,
128
+ "rstrip": false,
129
+ "single_word": false,
130
+ "special": true
131
+ },
132
+ "16": {
133
+ "content": "<empty_output>",
134
+ "lstrip": false,
135
+ "normalized": false,
136
+ "rstrip": false,
137
+ "single_word": false,
138
+ "special": true
139
+ }
140
+ },
141
+ "additional_special_tokens": [
142
+ "<|im_start|>",
143
+ "<|im_end|>"
144
+ ],
145
+ "bos_token": "<|im_start|>",
146
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
147
+ "clean_up_tokenization_spaces": false,
148
+ "eos_token": "<|im_end|>",
149
+ "extra_special_tokens": {},
150
+ "model_max_length": 8192,
151
+ "pad_token": "<|im_end|>",
152
+ "tokenizer_class": "GPT2Tokenizer",
153
+ "unk_token": "<|endoftext|>",
154
+ "vocab_size": 49152
155
+ }
checkpoint-2240/trainer_state.json ADDED
@@ -0,0 +1,2080 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.993312527864467,
6
+ "eval_steps": 500,
7
+ "global_step": 2240,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017833259028087384,
14
+ "grad_norm": 33.452999114990234,
15
+ "learning_rate": 3.2142857142857143e-06,
16
+ "loss": 8.9944,
17
+ "mean_token_accuracy": 0.32701094649964946,
18
+ "num_tokens": 2621440.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.03566651805617477,
23
+ "grad_norm": 2.596579074859619,
24
+ "learning_rate": 6.785714285714285e-06,
25
+ "loss": 4.9649,
26
+ "mean_token_accuracy": 0.5848119974136352,
27
+ "num_tokens": 5242880.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.05349977708426215,
32
+ "grad_norm": 1.8979898691177368,
33
+ "learning_rate": 1.0357142857142859e-05,
34
+ "loss": 4.1262,
35
+ "mean_token_accuracy": 0.5906830102205276,
36
+ "num_tokens": 7864320.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.07133303611234953,
41
+ "grad_norm": 1.1842058897018433,
42
+ "learning_rate": 1.392857142857143e-05,
43
+ "loss": 3.7327,
44
+ "mean_token_accuracy": 0.608327355235815,
45
+ "num_tokens": 10485760.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.08916629514043692,
50
+ "grad_norm": 0.922074556350708,
51
+ "learning_rate": 1.75e-05,
52
+ "loss": 3.4492,
53
+ "mean_token_accuracy": 0.6281662203371525,
54
+ "num_tokens": 13107200.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.1069995541685243,
59
+ "grad_norm": 0.7813541889190674,
60
+ "learning_rate": 2.107142857142857e-05,
61
+ "loss": 3.1416,
62
+ "mean_token_accuracy": 0.648241326212883,
63
+ "num_tokens": 15728640.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.12483281319661169,
68
+ "grad_norm": 1.1468125581741333,
69
+ "learning_rate": 2.4642857142857145e-05,
70
+ "loss": 2.8778,
71
+ "mean_token_accuracy": 0.6583510167896748,
72
+ "num_tokens": 18350080.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.14266607222469907,
77
+ "grad_norm": 1.0490431785583496,
78
+ "learning_rate": 2.8214285714285714e-05,
79
+ "loss": 2.7108,
80
+ "mean_token_accuracy": 0.663086223602295,
81
+ "num_tokens": 20971520.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.16049933125278645,
86
+ "grad_norm": 0.7813755869865417,
87
+ "learning_rate": 3.1785714285714284e-05,
88
+ "loss": 2.6039,
89
+ "mean_token_accuracy": 0.6650082431733608,
90
+ "num_tokens": 23592960.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.17833259028087384,
95
+ "grad_norm": 0.6500654220581055,
96
+ "learning_rate": 3.5357142857142864e-05,
97
+ "loss": 2.4513,
98
+ "mean_token_accuracy": 0.674950385093689,
99
+ "num_tokens": 26214400.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.19616584930896122,
104
+ "grad_norm": 0.81620192527771,
105
+ "learning_rate": 3.892857142857143e-05,
106
+ "loss": 2.3602,
107
+ "mean_token_accuracy": 0.6777540303766727,
108
+ "num_tokens": 28835840.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.2139991083370486,
113
+ "grad_norm": 0.6131238341331482,
114
+ "learning_rate": 4.25e-05,
115
+ "loss": 2.3545,
116
+ "mean_token_accuracy": 0.6758881151676178,
117
+ "num_tokens": 31457280.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.231832367365136,
122
+ "grad_norm": 0.9951689839363098,
123
+ "learning_rate": 4.607142857142857e-05,
124
+ "loss": 2.2624,
125
+ "mean_token_accuracy": 0.6806462205946445,
126
+ "num_tokens": 34078720.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.24966562639322337,
131
+ "grad_norm": 0.7796569466590881,
132
+ "learning_rate": 4.964285714285715e-05,
133
+ "loss": 2.1505,
134
+ "mean_token_accuracy": 0.6918394304811954,
135
+ "num_tokens": 36700160.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.26749888542131073,
140
+ "grad_norm": 0.8614610433578491,
141
+ "learning_rate": 4.9998587698517684e-05,
142
+ "loss": 2.1591,
143
+ "mean_token_accuracy": 0.68697941750288,
144
+ "num_tokens": 39321600.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.28533214444939814,
149
+ "grad_norm": 0.6178464889526367,
150
+ "learning_rate": 4.9993705873562665e-05,
151
+ "loss": 2.1048,
152
+ "mean_token_accuracy": 0.6911555036902428,
153
+ "num_tokens": 41943040.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.3031654034774855,
158
+ "grad_norm": 0.7255717515945435,
159
+ "learning_rate": 4.998533777009954e-05,
160
+ "loss": 2.0847,
161
+ "mean_token_accuracy": 0.6920825429260731,
162
+ "num_tokens": 44564480.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.3209986625055729,
167
+ "grad_norm": 0.7800336480140686,
168
+ "learning_rate": 4.9973484555363726e-05,
169
+ "loss": 2.049,
170
+ "mean_token_accuracy": 0.6982546880841255,
171
+ "num_tokens": 47185920.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.33883192153366026,
176
+ "grad_norm": 0.886745274066925,
177
+ "learning_rate": 4.995814788271582e-05,
178
+ "loss": 2.0301,
179
+ "mean_token_accuracy": 0.697366189956665,
180
+ "num_tokens": 49807360.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.3566651805617477,
185
+ "grad_norm": 0.6190801858901978,
186
+ "learning_rate": 4.993932989141088e-05,
187
+ "loss": 1.9408,
188
+ "mean_token_accuracy": 0.7095360569655895,
189
+ "num_tokens": 52428800.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.37449843958983503,
194
+ "grad_norm": 0.6333149671554565,
195
+ "learning_rate": 4.991703320630011e-05,
196
+ "loss": 1.9076,
197
+ "mean_token_accuracy": 0.7127362459897995,
198
+ "num_tokens": 55050240.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.39233169861792244,
203
+ "grad_norm": 0.9046627879142761,
204
+ "learning_rate": 4.989126093746468e-05,
205
+ "loss": 1.8913,
206
+ "mean_token_accuracy": 0.7148448958992958,
207
+ "num_tokens": 57671680.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.4101649576460098,
212
+ "grad_norm": 0.7778757214546204,
213
+ "learning_rate": 4.986201667978193e-05,
214
+ "loss": 1.9016,
215
+ "mean_token_accuracy": 0.7118340134620667,
216
+ "num_tokens": 60293120.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.4279982166740972,
221
+ "grad_norm": 0.6346383690834045,
222
+ "learning_rate": 4.9829304512423944e-05,
223
+ "loss": 1.898,
224
+ "mean_token_accuracy": 0.7123855009675026,
225
+ "num_tokens": 62914560.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.44583147570218457,
230
+ "grad_norm": 0.7308981418609619,
231
+ "learning_rate": 4.979312899828853e-05,
232
+ "loss": 1.8435,
233
+ "mean_token_accuracy": 0.7184088006615639,
234
+ "num_tokens": 65536000.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.463664734730272,
239
+ "grad_norm": 0.6174861788749695,
240
+ "learning_rate": 4.9753495183362796e-05,
241
+ "loss": 1.8242,
242
+ "mean_token_accuracy": 0.7238672457635402,
243
+ "num_tokens": 68157440.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.48149799375835933,
248
+ "grad_norm": 0.6507575511932373,
249
+ "learning_rate": 4.971040859601927e-05,
250
+ "loss": 1.791,
251
+ "mean_token_accuracy": 0.7251060992479325,
252
+ "num_tokens": 70778880.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.49933125278644674,
257
+ "grad_norm": 0.8092382550239563,
258
+ "learning_rate": 4.966387524624482e-05,
259
+ "loss": 1.7879,
260
+ "mean_token_accuracy": 0.7255102708935738,
261
+ "num_tokens": 73400320.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.5171645118145342,
266
+ "grad_norm": 0.5636172294616699,
267
+ "learning_rate": 4.9613901624802245e-05,
268
+ "loss": 1.7351,
269
+ "mean_token_accuracy": 0.7330663032829762,
270
+ "num_tokens": 76021760.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.5349977708426215,
275
+ "grad_norm": 0.655262291431427,
276
+ "learning_rate": 4.956049470232503e-05,
277
+ "loss": 1.7464,
278
+ "mean_token_accuracy": 0.7302443347871304,
279
+ "num_tokens": 78643200.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.5528310298707089,
284
+ "grad_norm": 0.9413332343101501,
285
+ "learning_rate": 4.9503661928344944e-05,
286
+ "loss": 1.7277,
287
+ "mean_token_accuracy": 0.730958030372858,
288
+ "num_tokens": 81264640.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.5706642888987963,
293
+ "grad_norm": 0.7439999580383301,
294
+ "learning_rate": 4.9443411230252956e-05,
295
+ "loss": 1.6947,
296
+ "mean_token_accuracy": 0.7346593342721462,
297
+ "num_tokens": 83886080.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.5884975479268837,
302
+ "grad_norm": 0.7665299773216248,
303
+ "learning_rate": 4.93797510121935e-05,
304
+ "loss": 1.7136,
305
+ "mean_token_accuracy": 0.7324949607253075,
306
+ "num_tokens": 86507520.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.606330806954971,
311
+ "grad_norm": 0.7366372346878052,
312
+ "learning_rate": 4.931269015389217e-05,
313
+ "loss": 1.676,
314
+ "mean_token_accuracy": 0.7371297933161258,
315
+ "num_tokens": 89128960.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.6241640659830584,
320
+ "grad_norm": 0.6571912169456482,
321
+ "learning_rate": 4.9242238009417175e-05,
322
+ "loss": 1.6252,
323
+ "mean_token_accuracy": 0.7434912383556366,
324
+ "num_tokens": 91750400.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.6419973250111458,
329
+ "grad_norm": 0.7496657967567444,
330
+ "learning_rate": 4.9168404405874505e-05,
331
+ "loss": 1.657,
332
+ "mean_token_accuracy": 0.737925547361374,
333
+ "num_tokens": 94371840.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.6598305840392331,
338
+ "grad_norm": 0.5275835394859314,
339
+ "learning_rate": 4.9091199642037244e-05,
340
+ "loss": 1.6723,
341
+ "mean_token_accuracy": 0.7358783438801766,
342
+ "num_tokens": 96993280.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.6776638430673205,
347
+ "grad_norm": 0.6143308281898499,
348
+ "learning_rate": 4.901063448690901e-05,
349
+ "loss": 1.6292,
350
+ "mean_token_accuracy": 0.7412570238113403,
351
+ "num_tokens": 99614720.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.6954971020954079,
356
+ "grad_norm": 0.860007643699646,
357
+ "learning_rate": 4.892672017822183e-05,
358
+ "loss": 1.6281,
359
+ "mean_token_accuracy": 0.7422710858285427,
360
+ "num_tokens": 102236160.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.7133303611234953,
365
+ "grad_norm": 0.5295814275741577,
366
+ "learning_rate": 4.8839468420868606e-05,
367
+ "loss": 1.6112,
368
+ "mean_token_accuracy": 0.7446457460522652,
369
+ "num_tokens": 104857600.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.7311636201515826,
374
+ "grad_norm": 0.7366165518760681,
375
+ "learning_rate": 4.8748891385270495e-05,
376
+ "loss": 1.6197,
377
+ "mean_token_accuracy": 0.7417318046092987,
378
+ "num_tokens": 107479040.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.7489968791796701,
383
+ "grad_norm": 0.7033573985099792,
384
+ "learning_rate": 4.865500170567926e-05,
385
+ "loss": 1.5763,
386
+ "mean_token_accuracy": 0.7474127545952797,
387
+ "num_tokens": 110100480.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.7668301382077575,
392
+ "grad_norm": 0.6527330875396729,
393
+ "learning_rate": 4.855781247841498e-05,
394
+ "loss": 1.5714,
395
+ "mean_token_accuracy": 0.7477886863052845,
396
+ "num_tokens": 112721920.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.7846633972358449,
401
+ "grad_norm": 0.7344551682472229,
402
+ "learning_rate": 4.845733726003928e-05,
403
+ "loss": 1.5682,
404
+ "mean_token_accuracy": 0.7475516766309738,
405
+ "num_tokens": 115343360.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.8024966562639322,
410
+ "grad_norm": 0.6840972304344177,
411
+ "learning_rate": 4.835359006546443e-05,
412
+ "loss": 1.5955,
413
+ "mean_token_accuracy": 0.7427977688610554,
414
+ "num_tokens": 117964800.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.8203299152920196,
419
+ "grad_norm": 0.635522723197937,
420
+ "learning_rate": 4.824658536599836e-05,
421
+ "loss": 1.5459,
422
+ "mean_token_accuracy": 0.7526196874678135,
423
+ "num_tokens": 120586240.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.838163174320107,
428
+ "grad_norm": 0.6874827146530151,
429
+ "learning_rate": 4.8136338087326216e-05,
430
+ "loss": 1.6108,
431
+ "mean_token_accuracy": 0.7413325920701027,
432
+ "num_tokens": 123207680.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.8559964333481944,
437
+ "grad_norm": 0.6571274995803833,
438
+ "learning_rate": 4.8022863607428356e-05,
439
+ "loss": 1.539,
440
+ "mean_token_accuracy": 0.7521994858980179,
441
+ "num_tokens": 125829120.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.8738296923762817,
446
+ "grad_norm": 0.8841336369514465,
447
+ "learning_rate": 4.790617775443537e-05,
448
+ "loss": 1.5486,
449
+ "mean_token_accuracy": 0.7492702730000019,
450
+ "num_tokens": 128450560.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.8916629514043691,
455
+ "grad_norm": 0.690990149974823,
456
+ "learning_rate": 4.7786296804420274e-05,
457
+ "loss": 1.5646,
458
+ "mean_token_accuracy": 0.7455468386411667,
459
+ "num_tokens": 131072000.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.9094962104324565,
464
+ "grad_norm": 0.7215222716331482,
465
+ "learning_rate": 4.766323747912818e-05,
466
+ "loss": 1.5085,
467
+ "mean_token_accuracy": 0.754824897646904,
468
+ "num_tokens": 133693440.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.927329469460544,
473
+ "grad_norm": 0.6024391651153564,
474
+ "learning_rate": 4.753701694364392e-05,
475
+ "loss": 1.4856,
476
+ "mean_token_accuracy": 0.7570274345576763,
477
+ "num_tokens": 136314880.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.9451627284886313,
482
+ "grad_norm": 0.6131455302238464,
483
+ "learning_rate": 4.740765280399767e-05,
484
+ "loss": 1.5433,
485
+ "mean_token_accuracy": 0.7495988786220551,
486
+ "num_tokens": 138936320.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.9629959875167187,
491
+ "grad_norm": 0.7875815629959106,
492
+ "learning_rate": 4.72751631047092e-05,
493
+ "loss": 1.4917,
494
+ "mean_token_accuracy": 0.7559954360127449,
495
+ "num_tokens": 141557760.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.9808292465448061,
500
+ "grad_norm": 0.7112794518470764,
501
+ "learning_rate": 4.71395663262709e-05,
502
+ "loss": 1.4784,
503
+ "mean_token_accuracy": 0.7572461232542992,
504
+ "num_tokens": 144179200.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.9986625055728935,
509
+ "grad_norm": 0.6807007789611816,
510
+ "learning_rate": 4.7000881382570014e-05,
511
+ "loss": 1.5159,
512
+ "mean_token_accuracy": 0.7523536741733551,
513
+ "num_tokens": 146800640.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 1.0,
518
+ "eval_loss": 1.494895339012146,
519
+ "eval_mean_token_accuracy": 0.7405715703964233,
520
+ "eval_num_tokens": 146980864.0,
521
+ "eval_runtime": 2.6056,
522
+ "eval_samples_per_second": 27.633,
523
+ "eval_steps_per_second": 1.919,
524
+ "step": 561
525
+ },
526
+ {
527
+ "epoch": 1.0160499331252786,
528
+ "grad_norm": 0.5480894446372986,
529
+ "learning_rate": 4.685912761825038e-05,
530
+ "loss": 1.4428,
531
+ "mean_token_accuracy": 0.7618689170250525,
532
+ "num_tokens": 149340160.0,
533
+ "step": 570
534
+ },
535
+ {
536
+ "epoch": 1.033883192153366,
537
+ "grad_norm": 0.6192435622215271,
538
+ "learning_rate": 4.671432480601416e-05,
539
+ "loss": 1.4711,
540
+ "mean_token_accuracy": 0.7556988887488842,
541
+ "num_tokens": 151961600.0,
542
+ "step": 580
543
+ },
544
+ {
545
+ "epoch": 1.0517164511814534,
546
+ "grad_norm": 0.5988848209381104,
547
+ "learning_rate": 4.656649314386378e-05,
548
+ "loss": 1.433,
549
+ "mean_token_accuracy": 0.7611474089324475,
550
+ "num_tokens": 154583040.0,
551
+ "step": 590
552
+ },
553
+ {
554
+ "epoch": 1.0695497102095408,
555
+ "grad_norm": 0.7051531672477722,
556
+ "learning_rate": 4.641565325228464e-05,
557
+ "loss": 1.415,
558
+ "mean_token_accuracy": 0.7643376767635346,
559
+ "num_tokens": 157204480.0,
560
+ "step": 600
561
+ },
562
+ {
563
+ "epoch": 1.0873829692376282,
564
+ "grad_norm": 0.6170579791069031,
565
+ "learning_rate": 4.6261826171368774e-05,
566
+ "loss": 1.4674,
567
+ "mean_token_accuracy": 0.7571297205984593,
568
+ "num_tokens": 159825920.0,
569
+ "step": 610
570
+ },
571
+ {
572
+ "epoch": 1.1052162282657156,
573
+ "grad_norm": 0.5691844820976257,
574
+ "learning_rate": 4.6105033357880104e-05,
575
+ "loss": 1.407,
576
+ "mean_token_accuracy": 0.7646078862249851,
577
+ "num_tokens": 162447360.0,
578
+ "step": 620
579
+ },
580
+ {
581
+ "epoch": 1.123049487293803,
582
+ "grad_norm": 0.5824171900749207,
583
+ "learning_rate": 4.594529668226146e-05,
584
+ "loss": 1.4185,
585
+ "mean_token_accuracy": 0.7614389970898628,
586
+ "num_tokens": 165068800.0,
587
+ "step": 630
588
+ },
589
+ {
590
+ "epoch": 1.1408827463218902,
591
+ "grad_norm": 0.5765758156776428,
592
+ "learning_rate": 4.578263842558402e-05,
593
+ "loss": 1.42,
594
+ "mean_token_accuracy": 0.7633033894002438,
595
+ "num_tokens": 167690240.0,
596
+ "step": 640
597
+ },
598
+ {
599
+ "epoch": 1.1587160053499777,
600
+ "grad_norm": 0.5261256694793701,
601
+ "learning_rate": 4.5617081276439355e-05,
602
+ "loss": 1.4384,
603
+ "mean_token_accuracy": 0.7604428693652153,
604
+ "num_tokens": 170311680.0,
605
+ "step": 650
606
+ },
607
+ {
608
+ "epoch": 1.176549264378065,
609
+ "grad_norm": 0.913208544254303,
610
+ "learning_rate": 4.544864832777469e-05,
611
+ "loss": 1.4775,
612
+ "mean_token_accuracy": 0.754877183586359,
613
+ "num_tokens": 172933120.0,
614
+ "step": 660
615
+ },
616
+ {
617
+ "epoch": 1.1943825234061525,
618
+ "grad_norm": 0.5799976587295532,
619
+ "learning_rate": 4.527736307367179e-05,
620
+ "loss": 1.4387,
621
+ "mean_token_accuracy": 0.7589173935353756,
622
+ "num_tokens": 175554560.0,
623
+ "step": 670
624
+ },
625
+ {
626
+ "epoch": 1.2122157824342399,
627
+ "grad_norm": 0.635047435760498,
628
+ "learning_rate": 4.510324940606979e-05,
629
+ "loss": 1.4439,
630
+ "mean_token_accuracy": 0.7597284145653248,
631
+ "num_tokens": 178176000.0,
632
+ "step": 680
633
+ },
634
+ {
635
+ "epoch": 1.2300490414623273,
636
+ "grad_norm": 0.5884873270988464,
637
+ "learning_rate": 4.4926331611432687e-05,
638
+ "loss": 1.3835,
639
+ "mean_token_accuracy": 0.7659787967801094,
640
+ "num_tokens": 180797440.0,
641
+ "step": 690
642
+ },
643
+ {
644
+ "epoch": 1.2478823004904147,
645
+ "grad_norm": 0.7232898473739624,
646
+ "learning_rate": 4.4746634367361634e-05,
647
+ "loss": 1.3629,
648
+ "mean_token_accuracy": 0.7691854700446129,
649
+ "num_tokens": 183418880.0,
650
+ "step": 700
651
+ },
652
+ {
653
+ "epoch": 1.265715559518502,
654
+ "grad_norm": 0.6569238305091858,
655
+ "learning_rate": 4.4564182739152805e-05,
656
+ "loss": 1.397,
657
+ "mean_token_accuracy": 0.7651830442249775,
658
+ "num_tokens": 186040320.0,
659
+ "step": 710
660
+ },
661
+ {
662
+ "epoch": 1.2835488185465893,
663
+ "grad_norm": 0.8630965352058411,
664
+ "learning_rate": 4.437900217630109e-05,
665
+ "loss": 1.4256,
666
+ "mean_token_accuracy": 0.7605882868170738,
667
+ "num_tokens": 188661760.0,
668
+ "step": 720
669
+ },
670
+ {
671
+ "epoch": 1.3013820775746767,
672
+ "grad_norm": 0.7600955367088318,
673
+ "learning_rate": 4.419111850895028e-05,
674
+ "loss": 1.3975,
675
+ "mean_token_accuracy": 0.7634797111153603,
676
+ "num_tokens": 191283200.0,
677
+ "step": 730
678
+ },
679
+ {
680
+ "epoch": 1.3192153366027641,
681
+ "grad_norm": 0.6687177419662476,
682
+ "learning_rate": 4.4000557944290086e-05,
683
+ "loss": 1.3616,
684
+ "mean_token_accuracy": 0.7703544825315476,
685
+ "num_tokens": 193904640.0,
686
+ "step": 740
687
+ },
688
+ {
689
+ "epoch": 1.3370485956308515,
690
+ "grad_norm": 0.719222903251648,
691
+ "learning_rate": 4.3807347062900624e-05,
692
+ "loss": 1.3788,
693
+ "mean_token_accuracy": 0.7648544363677502,
694
+ "num_tokens": 196526080.0,
695
+ "step": 750
696
+ },
697
+ {
698
+ "epoch": 1.354881854658939,
699
+ "grad_norm": 0.5986538529396057,
700
+ "learning_rate": 4.361151281504474e-05,
701
+ "loss": 1.364,
702
+ "mean_token_accuracy": 0.770479665696621,
703
+ "num_tokens": 199147520.0,
704
+ "step": 760
705
+ },
706
+ {
707
+ "epoch": 1.3727151136870264,
708
+ "grad_norm": 0.6472992897033691,
709
+ "learning_rate": 4.3413082516908893e-05,
710
+ "loss": 1.4045,
711
+ "mean_token_accuracy": 0.7644098080694676,
712
+ "num_tokens": 201768960.0,
713
+ "step": 770
714
+ },
715
+ {
716
+ "epoch": 1.3905483727151138,
717
+ "grad_norm": 0.44522207975387573,
718
+ "learning_rate": 4.321208384679285e-05,
719
+ "loss": 1.3314,
720
+ "mean_token_accuracy": 0.773588253557682,
721
+ "num_tokens": 204390400.0,
722
+ "step": 780
723
+ },
724
+ {
725
+ "epoch": 1.4083816317432012,
726
+ "grad_norm": 0.5019171833992004,
727
+ "learning_rate": 4.3008544841248986e-05,
728
+ "loss": 1.3591,
729
+ "mean_token_accuracy": 0.7695087313652038,
730
+ "num_tokens": 207011840.0,
731
+ "step": 790
732
+ },
733
+ {
734
+ "epoch": 1.4262148907712884,
735
+ "grad_norm": 0.7287567257881165,
736
+ "learning_rate": 4.280249389117157e-05,
737
+ "loss": 1.3773,
738
+ "mean_token_accuracy": 0.7675351880490779,
739
+ "num_tokens": 209633280.0,
740
+ "step": 800
741
+ },
742
+ {
743
+ "epoch": 1.4440481497993758,
744
+ "grad_norm": 1.0294618606567383,
745
+ "learning_rate": 4.2593959737836604e-05,
746
+ "loss": 1.3719,
747
+ "mean_token_accuracy": 0.7668005004525185,
748
+ "num_tokens": 212254720.0,
749
+ "step": 810
750
+ },
751
+ {
752
+ "epoch": 1.4618814088274632,
753
+ "grad_norm": 0.6181490421295166,
754
+ "learning_rate": 4.2382971468892806e-05,
755
+ "loss": 1.3168,
756
+ "mean_token_accuracy": 0.7758259050548076,
757
+ "num_tokens": 214876160.0,
758
+ "step": 820
759
+ },
760
+ {
761
+ "epoch": 1.4797146678555506,
762
+ "grad_norm": 0.5945810079574585,
763
+ "learning_rate": 4.216955851430432e-05,
764
+ "loss": 1.3536,
765
+ "mean_token_accuracy": 0.7710937492549419,
766
+ "num_tokens": 217497600.0,
767
+ "step": 830
768
+ },
769
+ {
770
+ "epoch": 1.497547926883638,
771
+ "grad_norm": 0.6803934574127197,
772
+ "learning_rate": 4.195375064224562e-05,
773
+ "loss": 1.324,
774
+ "mean_token_accuracy": 0.7753530308604241,
775
+ "num_tokens": 220119040.0,
776
+ "step": 840
777
+ },
778
+ {
779
+ "epoch": 1.5153811859117252,
780
+ "grad_norm": 0.5748502612113953,
781
+ "learning_rate": 4.1735577954949225e-05,
782
+ "loss": 1.3057,
783
+ "mean_token_accuracy": 0.7755480580031872,
784
+ "num_tokens": 222740480.0,
785
+ "step": 850
786
+ },
787
+ {
788
+ "epoch": 1.5332144449398126,
789
+ "grad_norm": 0.5675086379051208,
790
+ "learning_rate": 4.151507088450692e-05,
791
+ "loss": 1.302,
792
+ "mean_token_accuracy": 0.7765964694321156,
793
+ "num_tokens": 225361920.0,
794
+ "step": 860
795
+ },
796
+ {
797
+ "epoch": 1.5510477039679,
798
+ "grad_norm": 0.6266924142837524,
799
+ "learning_rate": 4.129226018862484e-05,
800
+ "loss": 1.353,
801
+ "mean_token_accuracy": 0.770804837346077,
802
+ "num_tokens": 227983360.0,
803
+ "step": 870
804
+ },
805
+ {
806
+ "epoch": 1.5688809629959874,
807
+ "grad_norm": 0.551527202129364,
808
+ "learning_rate": 4.106717694633321e-05,
809
+ "loss": 1.3477,
810
+ "mean_token_accuracy": 0.77206010222435,
811
+ "num_tokens": 230604800.0,
812
+ "step": 880
813
+ },
814
+ {
815
+ "epoch": 1.5867142220240749,
816
+ "grad_norm": 0.6628434658050537,
817
+ "learning_rate": 4.0839852553651265e-05,
818
+ "loss": 1.3244,
819
+ "mean_token_accuracy": 0.773486353456974,
820
+ "num_tokens": 233226240.0,
821
+ "step": 890
822
+ },
823
+ {
824
+ "epoch": 1.6045474810521623,
825
+ "grad_norm": 0.5956223011016846,
826
+ "learning_rate": 4.0610318719207906e-05,
827
+ "loss": 1.3298,
828
+ "mean_token_accuracy": 0.7733359761536122,
829
+ "num_tokens": 235847680.0,
830
+ "step": 900
831
+ },
832
+ {
833
+ "epoch": 1.6223807400802497,
834
+ "grad_norm": 0.6573621034622192,
835
+ "learning_rate": 4.037860745981881e-05,
836
+ "loss": 1.283,
837
+ "mean_token_accuracy": 0.7787142746150494,
838
+ "num_tokens": 238469120.0,
839
+ "step": 910
840
+ },
841
+ {
842
+ "epoch": 1.640213999108337,
843
+ "grad_norm": 0.607742965221405,
844
+ "learning_rate": 4.01447510960205e-05,
845
+ "loss": 1.2955,
846
+ "mean_token_accuracy": 0.7780937045812607,
847
+ "num_tokens": 241090560.0,
848
+ "step": 920
849
+ },
850
+ {
851
+ "epoch": 1.6580472581364245,
852
+ "grad_norm": 0.6278334856033325,
853
+ "learning_rate": 3.99087822475621e-05,
854
+ "loss": 1.3215,
855
+ "mean_token_accuracy": 0.7741382211446762,
856
+ "num_tokens": 243712000.0,
857
+ "step": 930
858
+ },
859
+ {
860
+ "epoch": 1.675880517164512,
861
+ "grad_norm": 0.5402811169624329,
862
+ "learning_rate": 3.96707338288553e-05,
863
+ "loss": 1.3113,
864
+ "mean_token_accuracy": 0.7751454092562199,
865
+ "num_tokens": 246333440.0,
866
+ "step": 940
867
+ },
868
+ {
869
+ "epoch": 1.6937137761925993,
870
+ "grad_norm": 0.6165571212768555,
871
+ "learning_rate": 3.9430639044383274e-05,
872
+ "loss": 1.3093,
873
+ "mean_token_accuracy": 0.7768628627061844,
874
+ "num_tokens": 248954880.0,
875
+ "step": 950
876
+ },
877
+ {
878
+ "epoch": 1.7115470352206867,
879
+ "grad_norm": 0.6086046695709229,
880
+ "learning_rate": 3.9188531384069096e-05,
881
+ "loss": 1.3125,
882
+ "mean_token_accuracy": 0.7762621335685254,
883
+ "num_tokens": 251576320.0,
884
+ "step": 960
885
+ },
886
+ {
887
+ "epoch": 1.729380294248774,
888
+ "grad_norm": 0.682686984539032,
889
+ "learning_rate": 3.8944444618604356e-05,
890
+ "loss": 1.2821,
891
+ "mean_token_accuracy": 0.7795726217329502,
892
+ "num_tokens": 254197760.0,
893
+ "step": 970
894
+ },
895
+ {
896
+ "epoch": 1.7472135532768613,
897
+ "grad_norm": 0.5615099668502808,
898
+ "learning_rate": 3.869841279473862e-05,
899
+ "loss": 1.3095,
900
+ "mean_token_accuracy": 0.7766918800771236,
901
+ "num_tokens": 256819200.0,
902
+ "step": 980
903
+ },
904
+ {
905
+ "epoch": 1.7650468123049488,
906
+ "grad_norm": 0.5341463088989258,
907
+ "learning_rate": 3.845047023053037e-05,
908
+ "loss": 1.2864,
909
+ "mean_token_accuracy": 0.779373399913311,
910
+ "num_tokens": 259440640.0,
911
+ "step": 990
912
+ },
913
+ {
914
+ "epoch": 1.7828800713330362,
915
+ "grad_norm": 0.5641157627105713,
916
+ "learning_rate": 3.82006515105601e-05,
917
+ "loss": 1.2857,
918
+ "mean_token_accuracy": 0.7798695519566536,
919
+ "num_tokens": 262062080.0,
920
+ "step": 1000
921
+ },
922
+ {
923
+ "epoch": 1.8007133303611234,
924
+ "grad_norm": 0.5196830630302429,
925
+ "learning_rate": 3.794899148110628e-05,
926
+ "loss": 1.2949,
927
+ "mean_token_accuracy": 0.7788909852504731,
928
+ "num_tokens": 264683520.0,
929
+ "step": 1010
930
+ },
931
+ {
932
+ "epoch": 1.8185465893892108,
933
+ "grad_norm": 0.4990008771419525,
934
+ "learning_rate": 3.7695525245284716e-05,
935
+ "loss": 1.254,
936
+ "mean_token_accuracy": 0.7839147254824639,
937
+ "num_tokens": 267304960.0,
938
+ "step": 1020
939
+ },
940
+ {
941
+ "epoch": 1.8363798484172982,
942
+ "grad_norm": 0.5643882751464844,
943
+ "learning_rate": 3.7440288158152187e-05,
944
+ "loss": 1.2446,
945
+ "mean_token_accuracy": 0.785726061463356,
946
+ "num_tokens": 269926400.0,
947
+ "step": 1030
948
+ },
949
+ {
950
+ "epoch": 1.8542131074453856,
951
+ "grad_norm": 0.8107313513755798,
952
+ "learning_rate": 3.7183315821774914e-05,
953
+ "loss": 1.2974,
954
+ "mean_token_accuracy": 0.7775464087724686,
955
+ "num_tokens": 272547840.0,
956
+ "step": 1040
957
+ },
958
+ {
959
+ "epoch": 1.872046366473473,
960
+ "grad_norm": 0.5176706314086914,
961
+ "learning_rate": 3.692464408026253e-05,
962
+ "loss": 1.2916,
963
+ "mean_token_accuracy": 0.7772151328623295,
964
+ "num_tokens": 275169280.0,
965
+ "step": 1050
966
+ },
967
+ {
968
+ "epoch": 1.8898796255015604,
969
+ "grad_norm": 0.5898165106773376,
970
+ "learning_rate": 3.666430901476833e-05,
971
+ "loss": 1.2716,
972
+ "mean_token_accuracy": 0.7806546173989772,
973
+ "num_tokens": 277790720.0,
974
+ "step": 1060
975
+ },
976
+ {
977
+ "epoch": 1.9077128845296478,
978
+ "grad_norm": 0.5368467569351196,
979
+ "learning_rate": 3.640234693845645e-05,
980
+ "loss": 1.2491,
981
+ "mean_token_accuracy": 0.7844746112823486,
982
+ "num_tokens": 280412160.0,
983
+ "step": 1070
984
+ },
985
+ {
986
+ "epoch": 1.9255461435577352,
987
+ "grad_norm": 0.5310569405555725,
988
+ "learning_rate": 3.613879439143667e-05,
989
+ "loss": 1.2825,
990
+ "mean_token_accuracy": 0.7795199528336525,
991
+ "num_tokens": 283033600.0,
992
+ "step": 1080
993
+ },
994
+ {
995
+ "epoch": 1.9433794025858226,
996
+ "grad_norm": 0.6114910840988159,
997
+ "learning_rate": 3.58736881356676e-05,
998
+ "loss": 1.2595,
999
+ "mean_token_accuracy": 0.7831220239400863,
1000
+ "num_tokens": 285655040.0,
1001
+ "step": 1090
1002
+ },
1003
+ {
1004
+ "epoch": 1.96121266161391,
1005
+ "grad_norm": 0.5786811113357544,
1006
+ "learning_rate": 3.5607065149828843e-05,
1007
+ "loss": 1.2694,
1008
+ "mean_token_accuracy": 0.7822808526456356,
1009
+ "num_tokens": 288276480.0,
1010
+ "step": 1100
1011
+ },
1012
+ {
1013
+ "epoch": 1.9790459206419975,
1014
+ "grad_norm": 0.6175279021263123,
1015
+ "learning_rate": 3.533896262416302e-05,
1016
+ "loss": 1.2712,
1017
+ "mean_token_accuracy": 0.7803981445729733,
1018
+ "num_tokens": 290897920.0,
1019
+ "step": 1110
1020
+ },
1021
+ {
1022
+ "epoch": 1.9968791796700847,
1023
+ "grad_norm": 0.568803608417511,
1024
+ "learning_rate": 3.506941795528821e-05,
1025
+ "loss": 1.2576,
1026
+ "mean_token_accuracy": 0.7821117818355561,
1027
+ "num_tokens": 293519360.0,
1028
+ "step": 1120
1029
+ },
1030
+ {
1031
+ "epoch": 2.0,
1032
+ "eval_loss": 1.2395988702774048,
1033
+ "eval_mean_token_accuracy": 0.7735649704933166,
1034
+ "eval_num_tokens": 293961728.0,
1035
+ "eval_runtime": 2.8804,
1036
+ "eval_samples_per_second": 24.997,
1037
+ "eval_steps_per_second": 1.736,
1038
+ "step": 1122
1039
+ },
1040
+ {
1041
+ "epoch": 2.0142666072224698,
1042
+ "grad_norm": 0.6652095913887024,
1043
+ "learning_rate": 3.479846874098167e-05,
1044
+ "loss": 1.2741,
1045
+ "mean_token_accuracy": 0.7794669828353784,
1046
+ "num_tokens": 296058880.0,
1047
+ "step": 1130
1048
+ },
1049
+ {
1050
+ "epoch": 2.032099866250557,
1051
+ "grad_norm": 0.7581707835197449,
1052
+ "learning_rate": 3.452615277493544e-05,
1053
+ "loss": 1.199,
1054
+ "mean_token_accuracy": 0.7900658771395683,
1055
+ "num_tokens": 298680320.0,
1056
+ "step": 1140
1057
+ },
1058
+ {
1059
+ "epoch": 2.0499331252786446,
1060
+ "grad_norm": 0.5383137464523315,
1061
+ "learning_rate": 3.425250804148464e-05,
1062
+ "loss": 1.2304,
1063
+ "mean_token_accuracy": 0.7861943542957306,
1064
+ "num_tokens": 301301760.0,
1065
+ "step": 1150
1066
+ },
1067
+ {
1068
+ "epoch": 2.067766384306732,
1069
+ "grad_norm": 0.5952672958374023,
1070
+ "learning_rate": 3.3977572710309166e-05,
1071
+ "loss": 1.2445,
1072
+ "mean_token_accuracy": 0.7836834371089936,
1073
+ "num_tokens": 303923200.0,
1074
+ "step": 1160
1075
+ },
1076
+ {
1077
+ "epoch": 2.0855996433348194,
1078
+ "grad_norm": 0.646511435508728,
1079
+ "learning_rate": 3.3701385131109616e-05,
1080
+ "loss": 1.2019,
1081
+ "mean_token_accuracy": 0.7892128698527813,
1082
+ "num_tokens": 306544640.0,
1083
+ "step": 1170
1084
+ },
1085
+ {
1086
+ "epoch": 2.103432902362907,
1087
+ "grad_norm": 0.514665961265564,
1088
+ "learning_rate": 3.342398382825791e-05,
1089
+ "loss": 1.2614,
1090
+ "mean_token_accuracy": 0.7812080189585686,
1091
+ "num_tokens": 309166080.0,
1092
+ "step": 1180
1093
+ },
1094
+ {
1095
+ "epoch": 2.121266161390994,
1096
+ "grad_norm": 0.5490918159484863,
1097
+ "learning_rate": 3.31454074954238e-05,
1098
+ "loss": 1.2108,
1099
+ "mean_token_accuracy": 0.7876022845506668,
1100
+ "num_tokens": 311787520.0,
1101
+ "step": 1190
1102
+ },
1103
+ {
1104
+ "epoch": 2.1390994204190816,
1105
+ "grad_norm": 0.6549789309501648,
1106
+ "learning_rate": 3.286569499017752e-05,
1107
+ "loss": 1.2024,
1108
+ "mean_token_accuracy": 0.7898712314665317,
1109
+ "num_tokens": 314408960.0,
1110
+ "step": 1200
1111
+ },
1112
+ {
1113
+ "epoch": 2.156932679447169,
1114
+ "grad_norm": 0.5743107199668884,
1115
+ "learning_rate": 3.2584885328569785e-05,
1116
+ "loss": 1.1835,
1117
+ "mean_token_accuracy": 0.7928275369107723,
1118
+ "num_tokens": 317030400.0,
1119
+ "step": 1210
1120
+ },
1121
+ {
1122
+ "epoch": 2.1747659384752565,
1123
+ "grad_norm": 0.6201761960983276,
1124
+ "learning_rate": 3.2303017679689525e-05,
1125
+ "loss": 1.1939,
1126
+ "mean_token_accuracy": 0.7908047571778297,
1127
+ "num_tokens": 319651840.0,
1128
+ "step": 1220
1129
+ },
1130
+ {
1131
+ "epoch": 2.192599197503344,
1132
+ "grad_norm": 0.550830602645874,
1133
+ "learning_rate": 3.2020131360200354e-05,
1134
+ "loss": 1.1864,
1135
+ "mean_token_accuracy": 0.7930233284831048,
1136
+ "num_tokens": 322273280.0,
1137
+ "step": 1230
1138
+ },
1139
+ {
1140
+ "epoch": 2.2104324565314313,
1141
+ "grad_norm": 0.5972590446472168,
1142
+ "learning_rate": 3.173626582885645e-05,
1143
+ "loss": 1.2177,
1144
+ "mean_token_accuracy": 0.7885610058903694,
1145
+ "num_tokens": 324894720.0,
1146
+ "step": 1240
1147
+ },
1148
+ {
1149
+ "epoch": 2.2282657155595187,
1150
+ "grad_norm": 0.5069220066070557,
1151
+ "learning_rate": 3.145146068099858e-05,
1152
+ "loss": 1.2279,
1153
+ "mean_token_accuracy": 0.7862870991230011,
1154
+ "num_tokens": 327516160.0,
1155
+ "step": 1250
1156
+ },
1157
+ {
1158
+ "epoch": 2.246098974587606,
1159
+ "grad_norm": 0.5100940465927124,
1160
+ "learning_rate": 3.116575564303109e-05,
1161
+ "loss": 1.1777,
1162
+ "mean_token_accuracy": 0.7929233327507973,
1163
+ "num_tokens": 330137600.0,
1164
+ "step": 1260
1165
+ },
1166
+ {
1167
+ "epoch": 2.2639322336156935,
1168
+ "grad_norm": 0.6175646781921387,
1169
+ "learning_rate": 3.087919056688067e-05,
1170
+ "loss": 1.173,
1171
+ "mean_token_accuracy": 0.7950430497527122,
1172
+ "num_tokens": 332759040.0,
1173
+ "step": 1270
1174
+ },
1175
+ {
1176
+ "epoch": 2.2817654926437805,
1177
+ "grad_norm": 0.5933928489685059,
1178
+ "learning_rate": 3.059180542443746e-05,
1179
+ "loss": 1.2057,
1180
+ "mean_token_accuracy": 0.7882316343486309,
1181
+ "num_tokens": 335380480.0,
1182
+ "step": 1280
1183
+ },
1184
+ {
1185
+ "epoch": 2.299598751671868,
1186
+ "grad_norm": 0.4882568120956421,
1187
+ "learning_rate": 3.0303640301979635e-05,
1188
+ "loss": 1.1982,
1189
+ "mean_token_accuracy": 0.7903318881988526,
1190
+ "num_tokens": 338001920.0,
1191
+ "step": 1290
1192
+ },
1193
+ {
1194
+ "epoch": 2.3174320106999553,
1195
+ "grad_norm": 0.5135630369186401,
1196
+ "learning_rate": 3.0014735394581823e-05,
1197
+ "loss": 1.2045,
1198
+ "mean_token_accuracy": 0.7893250770866871,
1199
+ "num_tokens": 340623360.0,
1200
+ "step": 1300
1201
+ },
1202
+ {
1203
+ "epoch": 2.3352652697280427,
1204
+ "grad_norm": 0.5496548414230347,
1205
+ "learning_rate": 2.972513100050851e-05,
1206
+ "loss": 1.1883,
1207
+ "mean_token_accuracy": 0.791755847632885,
1208
+ "num_tokens": 343244800.0,
1209
+ "step": 1310
1210
+ },
1211
+ {
1212
+ "epoch": 2.35309852875613,
1213
+ "grad_norm": 0.5601444840431213,
1214
+ "learning_rate": 2.943486751559295e-05,
1215
+ "loss": 1.1263,
1216
+ "mean_token_accuracy": 0.7998171858489513,
1217
+ "num_tokens": 345866240.0,
1218
+ "step": 1320
1219
+ },
1220
+ {
1221
+ "epoch": 2.3709317877842175,
1222
+ "grad_norm": 0.5243608355522156,
1223
+ "learning_rate": 2.9143985427602516e-05,
1224
+ "loss": 1.2274,
1225
+ "mean_token_accuracy": 0.7866500526666641,
1226
+ "num_tokens": 348487680.0,
1227
+ "step": 1330
1228
+ },
1229
+ {
1230
+ "epoch": 2.388765046812305,
1231
+ "grad_norm": 0.573567807674408,
1232
+ "learning_rate": 2.88525253105912e-05,
1233
+ "loss": 1.1977,
1234
+ "mean_token_accuracy": 0.7901960171759128,
1235
+ "num_tokens": 351109120.0,
1236
+ "step": 1340
1237
+ },
1238
+ {
1239
+ "epoch": 2.4065983058403924,
1240
+ "grad_norm": 0.4823901951313019,
1241
+ "learning_rate": 2.856052781924012e-05,
1242
+ "loss": 1.1358,
1243
+ "mean_token_accuracy": 0.7998057343065739,
1244
+ "num_tokens": 353730560.0,
1245
+ "step": 1350
1246
+ },
1247
+ {
1248
+ "epoch": 2.4244315648684798,
1249
+ "grad_norm": 0.515941858291626,
1250
+ "learning_rate": 2.8268033683186697e-05,
1251
+ "loss": 1.1881,
1252
+ "mean_token_accuracy": 0.7909879550337792,
1253
+ "num_tokens": 356352000.0,
1254
+ "step": 1360
1255
+ },
1256
+ {
1257
+ "epoch": 2.442264823896567,
1258
+ "grad_norm": 0.5508431196212769,
1259
+ "learning_rate": 2.7975083701343497e-05,
1260
+ "loss": 1.171,
1261
+ "mean_token_accuracy": 0.7945770494639873,
1262
+ "num_tokens": 358973440.0,
1263
+ "step": 1370
1264
+ },
1265
+ {
1266
+ "epoch": 2.4600980829246546,
1267
+ "grad_norm": 0.5473787188529968,
1268
+ "learning_rate": 2.7681718736207298e-05,
1269
+ "loss": 1.1309,
1270
+ "mean_token_accuracy": 0.7995454497635365,
1271
+ "num_tokens": 361594880.0,
1272
+ "step": 1380
1273
+ },
1274
+ {
1275
+ "epoch": 2.477931341952742,
1276
+ "grad_norm": 0.6680080890655518,
1277
+ "learning_rate": 2.738797970815933e-05,
1278
+ "loss": 1.1848,
1279
+ "mean_token_accuracy": 0.7917638584971428,
1280
+ "num_tokens": 364216320.0,
1281
+ "step": 1390
1282
+ },
1283
+ {
1284
+ "epoch": 2.4957646009808294,
1285
+ "grad_norm": 0.49958300590515137,
1286
+ "learning_rate": 2.7093907589757468e-05,
1287
+ "loss": 1.1922,
1288
+ "mean_token_accuracy": 0.7899307683110237,
1289
+ "num_tokens": 366837760.0,
1290
+ "step": 1400
1291
+ },
1292
+ {
1293
+ "epoch": 2.5135978600089164,
1294
+ "grad_norm": 0.4601416289806366,
1295
+ "learning_rate": 2.6799543400021132e-05,
1296
+ "loss": 1.1448,
1297
+ "mean_token_accuracy": 0.7981104202568531,
1298
+ "num_tokens": 369459200.0,
1299
+ "step": 1410
1300
+ },
1301
+ {
1302
+ "epoch": 2.531431119037004,
1303
+ "grad_norm": 0.5619862079620361,
1304
+ "learning_rate": 2.650492819870965e-05,
1305
+ "loss": 1.1371,
1306
+ "mean_token_accuracy": 0.7994878143072128,
1307
+ "num_tokens": 372080640.0,
1308
+ "step": 1420
1309
+ },
1310
+ {
1311
+ "epoch": 2.549264378065091,
1312
+ "grad_norm": 0.6030861139297485,
1313
+ "learning_rate": 2.621010308059504e-05,
1314
+ "loss": 1.1695,
1315
+ "mean_token_accuracy": 0.7944411806762218,
1316
+ "num_tokens": 374702080.0,
1317
+ "step": 1430
1318
+ },
1319
+ {
1320
+ "epoch": 2.5670976370931786,
1321
+ "grad_norm": 0.5237954258918762,
1322
+ "learning_rate": 2.5915109169729814e-05,
1323
+ "loss": 1.1372,
1324
+ "mean_token_accuracy": 0.7982218623161316,
1325
+ "num_tokens": 377323520.0,
1326
+ "step": 1440
1327
+ },
1328
+ {
1329
+ "epoch": 2.584930896121266,
1330
+ "grad_norm": 0.5144650936126709,
1331
+ "learning_rate": 2.5619987613710756e-05,
1332
+ "loss": 1.1896,
1333
+ "mean_token_accuracy": 0.7915455557405948,
1334
+ "num_tokens": 379944960.0,
1335
+ "step": 1450
1336
+ },
1337
+ {
1338
+ "epoch": 2.6027641551493534,
1339
+ "grad_norm": 0.5404797196388245,
1340
+ "learning_rate": 2.5324779577939394e-05,
1341
+ "loss": 1.1364,
1342
+ "mean_token_accuracy": 0.7987703062593937,
1343
+ "num_tokens": 382566400.0,
1344
+ "step": 1460
1345
+ },
1346
+ {
1347
+ "epoch": 2.620597414177441,
1348
+ "grad_norm": 0.4770708382129669,
1349
+ "learning_rate": 2.5029526239880015e-05,
1350
+ "loss": 1.1541,
1351
+ "mean_token_accuracy": 0.7973127588629723,
1352
+ "num_tokens": 385187840.0,
1353
+ "step": 1470
1354
+ },
1355
+ {
1356
+ "epoch": 2.6384306732055283,
1357
+ "grad_norm": 0.5579174757003784,
1358
+ "learning_rate": 2.473426878331591e-05,
1359
+ "loss": 1.1695,
1360
+ "mean_token_accuracy": 0.7945644512772561,
1361
+ "num_tokens": 387809280.0,
1362
+ "step": 1480
1363
+ },
1364
+ {
1365
+ "epoch": 2.6562639322336157,
1366
+ "grad_norm": 0.5287127494812012,
1367
+ "learning_rate": 2.443904839260488e-05,
1368
+ "loss": 1.1615,
1369
+ "mean_token_accuracy": 0.7949182473123073,
1370
+ "num_tokens": 390430720.0,
1371
+ "step": 1490
1372
+ },
1373
+ {
1374
+ "epoch": 2.674097191261703,
1375
+ "grad_norm": 0.5106226801872253,
1376
+ "learning_rate": 2.4143906246934537e-05,
1377
+ "loss": 1.1201,
1378
+ "mean_token_accuracy": 0.8007564425468445,
1379
+ "num_tokens": 393052160.0,
1380
+ "step": 1500
1381
+ },
1382
+ {
1383
+ "epoch": 2.6919304502897905,
1384
+ "grad_norm": 0.48907479643821716,
1385
+ "learning_rate": 2.384888351457838e-05,
1386
+ "loss": 1.1908,
1387
+ "mean_token_accuracy": 0.7911803111433983,
1388
+ "num_tokens": 395673600.0,
1389
+ "step": 1510
1390
+ },
1391
+ {
1392
+ "epoch": 2.709763709317878,
1393
+ "grad_norm": 0.5045855045318604,
1394
+ "learning_rate": 2.3554021347153403e-05,
1395
+ "loss": 1.1483,
1396
+ "mean_token_accuracy": 0.7968017213046551,
1397
+ "num_tokens": 398295040.0,
1398
+ "step": 1520
1399
+ },
1400
+ {
1401
+ "epoch": 2.7275969683459653,
1402
+ "grad_norm": 0.505662739276886,
1403
+ "learning_rate": 2.3259360873879992e-05,
1404
+ "loss": 1.1749,
1405
+ "mean_token_accuracy": 0.7930756144225597,
1406
+ "num_tokens": 400916480.0,
1407
+ "step": 1530
1408
+ },
1409
+ {
1410
+ "epoch": 2.7454302273740527,
1411
+ "grad_norm": 0.5283376574516296,
1412
+ "learning_rate": 2.2964943195844966e-05,
1413
+ "loss": 1.173,
1414
+ "mean_token_accuracy": 0.7921638391911984,
1415
+ "num_tokens": 403537920.0,
1416
+ "step": 1540
1417
+ },
1418
+ {
1419
+ "epoch": 2.76326348640214,
1420
+ "grad_norm": 0.5634790062904358,
1421
+ "learning_rate": 2.2670809380268548e-05,
1422
+ "loss": 1.1779,
1423
+ "mean_token_accuracy": 0.7936561174690724,
1424
+ "num_tokens": 406159360.0,
1425
+ "step": 1550
1426
+ },
1427
+ {
1428
+ "epoch": 2.7810967454302276,
1429
+ "grad_norm": 0.48228803277015686,
1430
+ "learning_rate": 2.237700045477607e-05,
1431
+ "loss": 1.14,
1432
+ "mean_token_accuracy": 0.796902097016573,
1433
+ "num_tokens": 408780800.0,
1434
+ "step": 1560
1435
+ },
1436
+ {
1437
+ "epoch": 2.798930004458315,
1438
+ "grad_norm": 0.4753395617008209,
1439
+ "learning_rate": 2.2083557401675115e-05,
1440
+ "loss": 1.1294,
1441
+ "mean_token_accuracy": 0.7987142026424408,
1442
+ "num_tokens": 411402240.0,
1443
+ "step": 1570
1444
+ },
1445
+ {
1446
+ "epoch": 2.8167632634864024,
1447
+ "grad_norm": 0.5348941683769226,
1448
+ "learning_rate": 2.1790521152239137e-05,
1449
+ "loss": 1.1581,
1450
+ "mean_token_accuracy": 0.7944865994155407,
1451
+ "num_tokens": 414023680.0,
1452
+ "step": 1580
1453
+ },
1454
+ {
1455
+ "epoch": 2.83459652251449,
1456
+ "grad_norm": 0.5610408782958984,
1457
+ "learning_rate": 2.1497932580998053e-05,
1458
+ "loss": 1.1609,
1459
+ "mean_token_accuracy": 0.795001070201397,
1460
+ "num_tokens": 416645120.0,
1461
+ "step": 1590
1462
+ },
1463
+ {
1464
+ "epoch": 2.8524297815425768,
1465
+ "grad_norm": 0.5431333184242249,
1466
+ "learning_rate": 2.1205832500036838e-05,
1467
+ "loss": 1.1405,
1468
+ "mean_token_accuracy": 0.7983539171516896,
1469
+ "num_tokens": 419266560.0,
1470
+ "step": 1600
1471
+ },
1472
+ {
1473
+ "epoch": 2.870263040570664,
1474
+ "grad_norm": 0.5650882720947266,
1475
+ "learning_rate": 2.0914261653302747e-05,
1476
+ "loss": 1.17,
1477
+ "mean_token_accuracy": 0.7925355695188046,
1478
+ "num_tokens": 421888000.0,
1479
+ "step": 1610
1480
+ },
1481
+ {
1482
+ "epoch": 2.8880962995987516,
1483
+ "grad_norm": 0.5435991287231445,
1484
+ "learning_rate": 2.062326071092218e-05,
1485
+ "loss": 1.1562,
1486
+ "mean_token_accuracy": 0.7965150982141495,
1487
+ "num_tokens": 424509440.0,
1488
+ "step": 1620
1489
+ },
1490
+ {
1491
+ "epoch": 2.905929558626839,
1492
+ "grad_norm": 0.4466145932674408,
1493
+ "learning_rate": 2.0332870263527675e-05,
1494
+ "loss": 1.1199,
1495
+ "mean_token_accuracy": 0.8016460791230202,
1496
+ "num_tokens": 427130880.0,
1497
+ "step": 1630
1498
+ },
1499
+ {
1500
+ "epoch": 2.9237628176549264,
1501
+ "grad_norm": 0.5083450675010681,
1502
+ "learning_rate": 2.0043130816596136e-05,
1503
+ "loss": 1.1467,
1504
+ "mean_token_accuracy": 0.7959990955889225,
1505
+ "num_tokens": 429752320.0,
1506
+ "step": 1640
1507
+ },
1508
+ {
1509
+ "epoch": 2.941596076683014,
1510
+ "grad_norm": 0.49085912108421326,
1511
+ "learning_rate": 1.9754082784798843e-05,
1512
+ "loss": 1.1565,
1513
+ "mean_token_accuracy": 0.7959273479878902,
1514
+ "num_tokens": 432373760.0,
1515
+ "step": 1650
1516
+ },
1517
+ {
1518
+ "epoch": 2.9594293357111012,
1519
+ "grad_norm": 0.48444733023643494,
1520
+ "learning_rate": 1.9465766486364143e-05,
1521
+ "loss": 1.1342,
1522
+ "mean_token_accuracy": 0.797958905249834,
1523
+ "num_tokens": 434995200.0,
1524
+ "step": 1660
1525
+ },
1526
+ {
1527
+ "epoch": 2.9772625947391886,
1528
+ "grad_norm": 0.5343450903892517,
1529
+ "learning_rate": 1.9178222137453656e-05,
1530
+ "loss": 1.141,
1531
+ "mean_token_accuracy": 0.7985142156481743,
1532
+ "num_tokens": 437616640.0,
1533
+ "step": 1670
1534
+ },
1535
+ {
1536
+ "epoch": 2.995095853767276,
1537
+ "grad_norm": 0.5583868026733398,
1538
+ "learning_rate": 1.8891489846552646e-05,
1539
+ "loss": 1.1432,
1540
+ "mean_token_accuracy": 0.7971971161663532,
1541
+ "num_tokens": 440238080.0,
1542
+ "step": 1680
1543
+ },
1544
+ {
1545
+ "epoch": 3.0,
1546
+ "eval_loss": 1.1234296560287476,
1547
+ "eval_mean_token_accuracy": 0.7894601821899414,
1548
+ "eval_num_tokens": 440942592.0,
1549
+ "eval_runtime": 2.9179,
1550
+ "eval_samples_per_second": 24.675,
1551
+ "eval_steps_per_second": 1.714,
1552
+ "step": 1683
1553
+ },
1554
+ {
1555
+ "epoch": 3.0124832813196614,
1556
+ "grad_norm": 0.5347409844398499,
1557
+ "learning_rate": 1.860560960887546e-05,
1558
+ "loss": 1.1253,
1559
+ "mean_token_accuracy": 0.8009015726737487,
1560
+ "num_tokens": 442777600.0,
1561
+ "step": 1690
1562
+ },
1563
+ {
1564
+ "epoch": 3.0303165403477483,
1565
+ "grad_norm": 0.5598998069763184,
1566
+ "learning_rate": 1.8320621300786688e-05,
1567
+ "loss": 1.106,
1568
+ "mean_token_accuracy": 0.8023651242256165,
1569
+ "num_tokens": 445399040.0,
1570
+ "step": 1700
1571
+ },
1572
+ {
1573
+ "epoch": 3.0481497993758357,
1574
+ "grad_norm": 0.5113139152526855,
1575
+ "learning_rate": 1.8036564674239043e-05,
1576
+ "loss": 1.0887,
1577
+ "mean_token_accuracy": 0.8040673099458218,
1578
+ "num_tokens": 448020480.0,
1579
+ "step": 1710
1580
+ },
1581
+ {
1582
+ "epoch": 3.065983058403923,
1583
+ "grad_norm": 0.44441869854927063,
1584
+ "learning_rate": 1.7753479351228444e-05,
1585
+ "loss": 1.1185,
1586
+ "mean_token_accuracy": 0.7999229028820991,
1587
+ "num_tokens": 450641920.0,
1588
+ "step": 1720
1589
+ },
1590
+ {
1591
+ "epoch": 3.0838163174320106,
1592
+ "grad_norm": 0.47339409589767456,
1593
+ "learning_rate": 1.7471404818267316e-05,
1594
+ "loss": 1.0925,
1595
+ "mean_token_accuracy": 0.8034070417284965,
1596
+ "num_tokens": 453263360.0,
1597
+ "step": 1730
1598
+ },
1599
+ {
1600
+ "epoch": 3.101649576460098,
1601
+ "grad_norm": 0.47856712341308594,
1602
+ "learning_rate": 1.7190380420876766e-05,
1603
+ "loss": 1.0911,
1604
+ "mean_token_accuracy": 0.8045188069343567,
1605
+ "num_tokens": 455884800.0,
1606
+ "step": 1740
1607
+ },
1608
+ {
1609
+ "epoch": 3.1194828354881854,
1610
+ "grad_norm": 0.4918275773525238,
1611
+ "learning_rate": 1.6910445358098415e-05,
1612
+ "loss": 1.1109,
1613
+ "mean_token_accuracy": 0.8015380732715129,
1614
+ "num_tokens": 458506240.0,
1615
+ "step": 1750
1616
+ },
1617
+ {
1618
+ "epoch": 3.137316094516273,
1619
+ "grad_norm": 0.5043804049491882,
1620
+ "learning_rate": 1.6631638677026674e-05,
1621
+ "loss": 1.1364,
1622
+ "mean_token_accuracy": 0.7974261090159416,
1623
+ "num_tokens": 461127680.0,
1624
+ "step": 1760
1625
+ },
1626
+ {
1627
+ "epoch": 3.15514935354436,
1628
+ "grad_norm": 0.4904772937297821,
1629
+ "learning_rate": 1.6353999267362237e-05,
1630
+ "loss": 1.0838,
1631
+ "mean_token_accuracy": 0.806386636942625,
1632
+ "num_tokens": 463749120.0,
1633
+ "step": 1770
1634
+ },
1635
+ {
1636
+ "epoch": 3.1729826125724476,
1637
+ "grad_norm": 0.48294153809547424,
1638
+ "learning_rate": 1.607756585598743e-05,
1639
+ "loss": 1.1389,
1640
+ "mean_token_accuracy": 0.7980314172804356,
1641
+ "num_tokens": 466370560.0,
1642
+ "step": 1780
1643
+ },
1644
+ {
1645
+ "epoch": 3.190815871600535,
1646
+ "grad_norm": 0.5139598250389099,
1647
+ "learning_rate": 1.580237700156443e-05,
1648
+ "loss": 1.1095,
1649
+ "mean_token_accuracy": 0.8025780878961086,
1650
+ "num_tokens": 468992000.0,
1651
+ "step": 1790
1652
+ },
1653
+ {
1654
+ "epoch": 3.2086491306286224,
1655
+ "grad_norm": 0.47298020124435425,
1656
+ "learning_rate": 1.5528471089156804e-05,
1657
+ "loss": 1.0931,
1658
+ "mean_token_accuracy": 0.8041585251688957,
1659
+ "num_tokens": 471613440.0,
1660
+ "step": 1800
1661
+ },
1662
+ {
1663
+ "epoch": 3.22648238965671,
1664
+ "grad_norm": 0.41903576254844666,
1665
+ "learning_rate": 1.5255886324875363e-05,
1666
+ "loss": 1.0954,
1667
+ "mean_token_accuracy": 0.8047371134161949,
1668
+ "num_tokens": 474234880.0,
1669
+ "step": 1810
1670
+ },
1671
+ {
1672
+ "epoch": 3.2443156486847973,
1673
+ "grad_norm": 0.4610045552253723,
1674
+ "learning_rate": 1.4984660730548856e-05,
1675
+ "loss": 1.1095,
1676
+ "mean_token_accuracy": 0.8022483348846435,
1677
+ "num_tokens": 476856320.0,
1678
+ "step": 1820
1679
+ },
1680
+ {
1681
+ "epoch": 3.2621489077128847,
1682
+ "grad_norm": 0.5105413198471069,
1683
+ "learning_rate": 1.4714832138420553e-05,
1684
+ "loss": 1.1258,
1685
+ "mean_token_accuracy": 0.7983997173607349,
1686
+ "num_tokens": 479477760.0,
1687
+ "step": 1830
1688
+ },
1689
+ {
1690
+ "epoch": 3.279982166740972,
1691
+ "grad_norm": 0.4393883943557739,
1692
+ "learning_rate": 1.4446438185871092e-05,
1693
+ "loss": 1.0566,
1694
+ "mean_token_accuracy": 0.8093036323785782,
1695
+ "num_tokens": 482099200.0,
1696
+ "step": 1840
1697
+ },
1698
+ {
1699
+ "epoch": 3.297815425769059,
1700
+ "grad_norm": 0.4618268311023712,
1701
+ "learning_rate": 1.4179516310168623e-05,
1702
+ "loss": 1.0697,
1703
+ "mean_token_accuracy": 0.8082189694046974,
1704
+ "num_tokens": 484720640.0,
1705
+ "step": 1850
1706
+ },
1707
+ {
1708
+ "epoch": 3.3156486847971465,
1709
+ "grad_norm": 0.47192856669425964,
1710
+ "learning_rate": 1.3914103743246757e-05,
1711
+ "loss": 1.1203,
1712
+ "mean_token_accuracy": 0.8002961635589599,
1713
+ "num_tokens": 487342080.0,
1714
+ "step": 1860
1715
+ },
1716
+ {
1717
+ "epoch": 3.333481943825234,
1718
+ "grad_norm": 0.4617849290370941,
1719
+ "learning_rate": 1.3650237506511331e-05,
1720
+ "loss": 1.0664,
1721
+ "mean_token_accuracy": 0.8089914366602897,
1722
+ "num_tokens": 489963520.0,
1723
+ "step": 1870
1724
+ },
1725
+ {
1726
+ "epoch": 3.3513152028533213,
1727
+ "grad_norm": 0.46263769268989563,
1728
+ "learning_rate": 1.3387954405676329e-05,
1729
+ "loss": 1.0855,
1730
+ "mean_token_accuracy": 0.8045512497425079,
1731
+ "num_tokens": 492584960.0,
1732
+ "step": 1880
1733
+ },
1734
+ {
1735
+ "epoch": 3.3691484618814087,
1736
+ "grad_norm": 0.47842440009117126,
1737
+ "learning_rate": 1.3127291025630073e-05,
1738
+ "loss": 1.0705,
1739
+ "mean_token_accuracy": 0.8085929870605468,
1740
+ "num_tokens": 495206400.0,
1741
+ "step": 1890
1742
+ },
1743
+ {
1744
+ "epoch": 3.386981720909496,
1745
+ "grad_norm": 0.43500861525535583,
1746
+ "learning_rate": 1.2868283725332028e-05,
1747
+ "loss": 1.0635,
1748
+ "mean_token_accuracy": 0.8091395162045956,
1749
+ "num_tokens": 497827840.0,
1750
+ "step": 1900
1751
+ },
1752
+ {
1753
+ "epoch": 3.4048149799375835,
1754
+ "grad_norm": 0.508797287940979,
1755
+ "learning_rate": 1.261096863274136e-05,
1756
+ "loss": 1.1052,
1757
+ "mean_token_accuracy": 0.8021300226449967,
1758
+ "num_tokens": 500449280.0,
1759
+ "step": 1910
1760
+ },
1761
+ {
1762
+ "epoch": 3.422648238965671,
1763
+ "grad_norm": 0.42378219962120056,
1764
+ "learning_rate": 1.2355381639777472e-05,
1765
+ "loss": 1.0936,
1766
+ "mean_token_accuracy": 0.8043497301638126,
1767
+ "num_tokens": 503070720.0,
1768
+ "step": 1920
1769
+ },
1770
+ {
1771
+ "epoch": 3.4404814979937584,
1772
+ "grad_norm": 0.4547331631183624,
1773
+ "learning_rate": 1.2101558397313634e-05,
1774
+ "loss": 1.1142,
1775
+ "mean_token_accuracy": 0.8015869230031967,
1776
+ "num_tokens": 505692160.0,
1777
+ "step": 1930
1778
+ },
1779
+ {
1780
+ "epoch": 3.4583147570218458,
1781
+ "grad_norm": 0.4319451153278351,
1782
+ "learning_rate": 1.1849534310204152e-05,
1783
+ "loss": 1.0835,
1784
+ "mean_token_accuracy": 0.8058400988578797,
1785
+ "num_tokens": 508313600.0,
1786
+ "step": 1940
1787
+ },
1788
+ {
1789
+ "epoch": 3.476148016049933,
1790
+ "grad_norm": 0.4666835069656372,
1791
+ "learning_rate": 1.1599344532345877e-05,
1792
+ "loss": 1.1072,
1793
+ "mean_token_accuracy": 0.801591893285513,
1794
+ "num_tokens": 510935040.0,
1795
+ "step": 1950
1796
+ },
1797
+ {
1798
+ "epoch": 3.4939812750780206,
1799
+ "grad_norm": 0.47252121567726135,
1800
+ "learning_rate": 1.1351023961774749e-05,
1801
+ "loss": 1.1245,
1802
+ "mean_token_accuracy": 0.7985180281102657,
1803
+ "num_tokens": 513556480.0,
1804
+ "step": 1960
1805
+ },
1806
+ {
1807
+ "epoch": 3.511814534106108,
1808
+ "grad_norm": 0.5152117013931274,
1809
+ "learning_rate": 1.110460723579796e-05,
1810
+ "loss": 1.0982,
1811
+ "mean_token_accuracy": 0.8033757455646992,
1812
+ "num_tokens": 516177920.0,
1813
+ "step": 1970
1814
+ },
1815
+ {
1816
+ "epoch": 3.5296477931341954,
1817
+ "grad_norm": 0.43357202410697937,
1818
+ "learning_rate": 1.0860128726162545e-05,
1819
+ "loss": 1.0632,
1820
+ "mean_token_accuracy": 0.8095081992447376,
1821
+ "num_tokens": 518799360.0,
1822
+ "step": 1980
1823
+ },
1824
+ {
1825
+ "epoch": 3.547481052162283,
1826
+ "grad_norm": 0.45368409156799316,
1827
+ "learning_rate": 1.0617622534260985e-05,
1828
+ "loss": 1.114,
1829
+ "mean_token_accuracy": 0.8014212884008884,
1830
+ "num_tokens": 521420800.0,
1831
+ "step": 1990
1832
+ },
1833
+ {
1834
+ "epoch": 3.5653143111903702,
1835
+ "grad_norm": 0.38939619064331055,
1836
+ "learning_rate": 1.0377122486374576e-05,
1837
+ "loss": 1.0449,
1838
+ "mean_token_accuracy": 0.8123023062944412,
1839
+ "num_tokens": 524042240.0,
1840
+ "step": 2000
1841
+ },
1842
+ {
1843
+ "epoch": 3.5831475702184576,
1844
+ "grad_norm": 0.42768579721450806,
1845
+ "learning_rate": 1.0138662128955053e-05,
1846
+ "loss": 1.104,
1847
+ "mean_token_accuracy": 0.8025334358215332,
1848
+ "num_tokens": 526663680.0,
1849
+ "step": 2010
1850
+ },
1851
+ {
1852
+ "epoch": 3.600980829246545,
1853
+ "grad_norm": 0.41337716579437256,
1854
+ "learning_rate": 9.902274723945381e-06,
1855
+ "loss": 1.0788,
1856
+ "mean_token_accuracy": 0.8076182380318642,
1857
+ "num_tokens": 529285120.0,
1858
+ "step": 2020
1859
+ },
1860
+ {
1861
+ "epoch": 3.618814088274632,
1862
+ "grad_norm": 0.4830559194087982,
1863
+ "learning_rate": 9.667993244140134e-06,
1864
+ "loss": 1.0967,
1865
+ "mean_token_accuracy": 0.8028597474098206,
1866
+ "num_tokens": 531906560.0,
1867
+ "step": 2030
1868
+ },
1869
+ {
1870
+ "epoch": 3.6366473473027194,
1871
+ "grad_norm": 0.4195765554904938,
1872
+ "learning_rate": 9.435850368586286e-06,
1873
+ "loss": 1.0773,
1874
+ "mean_token_accuracy": 0.8066488340497017,
1875
+ "num_tokens": 534528000.0,
1876
+ "step": 2040
1877
+ },
1878
+ {
1879
+ "epoch": 3.654480606330807,
1880
+ "grad_norm": 0.4251314103603363,
1881
+ "learning_rate": 9.20587847802488e-06,
1882
+ "loss": 1.0819,
1883
+ "mean_token_accuracy": 0.8052599862217903,
1884
+ "num_tokens": 537149440.0,
1885
+ "step": 2050
1886
+ },
1887
+ {
1888
+ "epoch": 3.6723138653588943,
1889
+ "grad_norm": 0.4114632308483124,
1890
+ "learning_rate": 8.978109650374397e-06,
1891
+ "loss": 1.0697,
1892
+ "mean_token_accuracy": 0.8077415138483047,
1893
+ "num_tokens": 539770880.0,
1894
+ "step": 2060
1895
+ },
1896
+ {
1897
+ "epoch": 3.6901471243869817,
1898
+ "grad_norm": 0.4304318428039551,
1899
+ "learning_rate": 8.752575656256321e-06,
1900
+ "loss": 1.1046,
1901
+ "mean_token_accuracy": 0.8016697481274605,
1902
+ "num_tokens": 542392320.0,
1903
+ "step": 2070
1904
+ },
1905
+ {
1906
+ "epoch": 3.707980383415069,
1907
+ "grad_norm": 0.47687673568725586,
1908
+ "learning_rate": 8.52930795456355e-06,
1909
+ "loss": 1.0558,
1910
+ "mean_token_accuracy": 0.8107806429266929,
1911
+ "num_tokens": 545013760.0,
1912
+ "step": 2080
1913
+ },
1914
+ {
1915
+ "epoch": 3.7258136424431565,
1916
+ "grad_norm": 0.45524051785469055,
1917
+ "learning_rate": 8.30833768807237e-06,
1918
+ "loss": 1.0816,
1919
+ "mean_token_accuracy": 0.8059305518865585,
1920
+ "num_tokens": 547635200.0,
1921
+ "step": 2090
1922
+ },
1923
+ {
1924
+ "epoch": 3.743646901471244,
1925
+ "grad_norm": 0.45478421449661255,
1926
+ "learning_rate": 8.089695679098375e-06,
1927
+ "loss": 1.0961,
1928
+ "mean_token_accuracy": 0.8029544025659561,
1929
+ "num_tokens": 550256640.0,
1930
+ "step": 2100
1931
+ },
1932
+ {
1933
+ "epoch": 3.7614801604993313,
1934
+ "grad_norm": 0.45910394191741943,
1935
+ "learning_rate": 7.87341242519728e-06,
1936
+ "loss": 1.0608,
1937
+ "mean_token_accuracy": 0.8082987323403359,
1938
+ "num_tokens": 552878080.0,
1939
+ "step": 2110
1940
+ },
1941
+ {
1942
+ "epoch": 3.7793134195274187,
1943
+ "grad_norm": 0.4360869824886322,
1944
+ "learning_rate": 7.659518094910839e-06,
1945
+ "loss": 1.0806,
1946
+ "mean_token_accuracy": 0.8052332706749439,
1947
+ "num_tokens": 555499520.0,
1948
+ "step": 2120
1949
+ },
1950
+ {
1951
+ "epoch": 3.797146678555506,
1952
+ "grad_norm": 0.46505963802337646,
1953
+ "learning_rate": 7.44804252355884e-06,
1954
+ "loss": 1.1143,
1955
+ "mean_token_accuracy": 0.8007854476571084,
1956
+ "num_tokens": 558120960.0,
1957
+ "step": 2130
1958
+ },
1959
+ {
1960
+ "epoch": 3.8149799375835935,
1961
+ "grad_norm": 0.40090060234069824,
1962
+ "learning_rate": 7.2390152090773765e-06,
1963
+ "loss": 1.079,
1964
+ "mean_token_accuracy": 0.8058068990707398,
1965
+ "num_tokens": 560742400.0,
1966
+ "step": 2140
1967
+ },
1968
+ {
1969
+ "epoch": 3.8328131966116805,
1970
+ "grad_norm": 0.41377872228622437,
1971
+ "learning_rate": 7.032465307904404e-06,
1972
+ "loss": 1.0496,
1973
+ "mean_token_accuracy": 0.8097444467246533,
1974
+ "num_tokens": 563363840.0,
1975
+ "step": 2150
1976
+ },
1977
+ {
1978
+ "epoch": 3.850646455639768,
1979
+ "grad_norm": 0.45236936211586,
1980
+ "learning_rate": 6.82842163091274e-06,
1981
+ "loss": 1.0819,
1982
+ "mean_token_accuracy": 0.8053412742912769,
1983
+ "num_tokens": 565985280.0,
1984
+ "step": 2160
1985
+ },
1986
+ {
1987
+ "epoch": 3.8684797146678553,
1988
+ "grad_norm": 0.42451992630958557,
1989
+ "learning_rate": 6.626912639391377e-06,
1990
+ "loss": 1.0649,
1991
+ "mean_token_accuracy": 0.8076308310031891,
1992
+ "num_tokens": 568606720.0,
1993
+ "step": 2170
1994
+ },
1995
+ {
1996
+ "epoch": 3.8863129736959428,
1997
+ "grad_norm": 0.43241724371910095,
1998
+ "learning_rate": 6.427966441075511e-06,
1999
+ "loss": 1.0821,
2000
+ "mean_token_accuracy": 0.805367611348629,
2001
+ "num_tokens": 571228160.0,
2002
+ "step": 2180
2003
+ },
2004
+ {
2005
+ "epoch": 3.90414623272403,
2006
+ "grad_norm": 0.41288912296295166,
2007
+ "learning_rate": 6.2316107862259026e-06,
2008
+ "loss": 1.1017,
2009
+ "mean_token_accuracy": 0.8022395603358745,
2010
+ "num_tokens": 573849600.0,
2011
+ "step": 2190
2012
+ },
2013
+ {
2014
+ "epoch": 3.9219794917521176,
2015
+ "grad_norm": 0.4336395263671875,
2016
+ "learning_rate": 6.037873063758112e-06,
2017
+ "loss": 1.0546,
2018
+ "mean_token_accuracy": 0.8097028434276581,
2019
+ "num_tokens": 576471040.0,
2020
+ "step": 2200
2021
+ },
2022
+ {
2023
+ "epoch": 3.939812750780205,
2024
+ "grad_norm": 0.4041893482208252,
2025
+ "learning_rate": 5.846780297422091e-06,
2026
+ "loss": 1.0603,
2027
+ "mean_token_accuracy": 0.8089009828865528,
2028
+ "num_tokens": 579092480.0,
2029
+ "step": 2210
2030
+ },
2031
+ {
2032
+ "epoch": 3.9576460098082924,
2033
+ "grad_norm": 0.39978188276290894,
2034
+ "learning_rate": 5.6583591420327684e-06,
2035
+ "loss": 1.0752,
2036
+ "mean_token_accuracy": 0.8069194242358207,
2037
+ "num_tokens": 581713920.0,
2038
+ "step": 2220
2039
+ },
2040
+ {
2041
+ "epoch": 3.97547926883638,
2042
+ "grad_norm": 0.41306570172309875,
2043
+ "learning_rate": 5.472635879752042e-06,
2044
+ "loss": 1.0757,
2045
+ "mean_token_accuracy": 0.8059477277100087,
2046
+ "num_tokens": 584335360.0,
2047
+ "step": 2230
2048
+ },
2049
+ {
2050
+ "epoch": 3.993312527864467,
2051
+ "grad_norm": 0.4413711428642273,
2052
+ "learning_rate": 5.289636416422822e-06,
2053
+ "loss": 1.1117,
2054
+ "mean_token_accuracy": 0.8009537614881992,
2055
+ "num_tokens": 586956800.0,
2056
+ "step": 2240
2057
+ }
2058
+ ],
2059
+ "logging_steps": 10,
2060
+ "max_steps": 2800,
2061
+ "num_input_tokens_seen": 0,
2062
+ "num_train_epochs": 5,
2063
+ "save_steps": 140,
2064
+ "stateful_callbacks": {
2065
+ "TrainerControl": {
2066
+ "args": {
2067
+ "should_epoch_stop": false,
2068
+ "should_evaluate": false,
2069
+ "should_log": false,
2070
+ "should_save": true,
2071
+ "should_training_stop": false
2072
+ },
2073
+ "attributes": {}
2074
+ }
2075
+ },
2076
+ "total_flos": 2.922512307965657e+18,
2077
+ "train_batch_size": 16,
2078
+ "trial_name": null,
2079
+ "trial_params": null
2080
+ }
checkpoint-2240/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7409b037df1da3374237680bbedeb80953c01e549dccfb71a127e04d8bc7415b
3
+ size 7352
checkpoint-2240/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2380/config.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_configs": [
3
+ {
4
+ "vocab_size": 49152,
5
+ "max_position_embeddings": 8192,
6
+ "hidden_size": 960,
7
+ "intermediate_size": 2560,
8
+ "num_hidden_layers": 32,
9
+ "num_attention_heads": 15,
10
+ "num_key_value_heads": 5,
11
+ "hidden_act": "silu",
12
+ "initializer_range": 0.02,
13
+ "rms_norm_eps": 1e-05,
14
+ "pretraining_tp": 1,
15
+ "use_cache": true,
16
+ "rope_theta": 100000,
17
+ "rope_scaling": null,
18
+ "attention_bias": false,
19
+ "attention_dropout": 0.0,
20
+ "mlp_bias": false,
21
+ "head_dim": 64,
22
+ "return_dict": true,
23
+ "output_hidden_states": false,
24
+ "output_attentions": false,
25
+ "torchscript": false,
26
+ "torch_dtype": "float32",
27
+ "use_bfloat16": false,
28
+ "tf_legacy_loss": false,
29
+ "pruned_heads": {},
30
+ "tie_word_embeddings": true,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "is_decoder": false,
34
+ "cross_attention_hidden_size": null,
35
+ "add_cross_attention": false,
36
+ "tie_encoder_decoder": false,
37
+ "max_length": 20,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "early_stopping": false,
41
+ "num_beams": 1,
42
+ "num_beam_groups": 1,
43
+ "diversity_penalty": 0.0,
44
+ "temperature": 1.0,
45
+ "top_k": 50,
46
+ "top_p": 1.0,
47
+ "typical_p": 1.0,
48
+ "repetition_penalty": 1.0,
49
+ "length_penalty": 1.0,
50
+ "no_repeat_ngram_size": 0,
51
+ "encoder_no_repeat_ngram_size": 0,
52
+ "bad_words_ids": null,
53
+ "num_return_sequences": 1,
54
+ "output_scores": false,
55
+ "return_dict_in_generate": false,
56
+ "forced_bos_token_id": null,
57
+ "forced_eos_token_id": null,
58
+ "remove_invalid_values": false,
59
+ "exponential_decay_length_penalty": null,
60
+ "suppress_tokens": null,
61
+ "begin_suppress_tokens": null,
62
+ "architectures": [
63
+ "LlamaForCausalLM"
64
+ ],
65
+ "finetuning_task": null,
66
+ "id2label": {
67
+ "0": "LABEL_0",
68
+ "1": "LABEL_1"
69
+ },
70
+ "label2id": {
71
+ "LABEL_0": 0,
72
+ "LABEL_1": 1
73
+ },
74
+ "tokenizer_class": null,
75
+ "prefix": null,
76
+ "bos_token_id": 1,
77
+ "pad_token_id": 2,
78
+ "eos_token_id": 2,
79
+ "sep_token_id": null,
80
+ "decoder_start_token_id": null,
81
+ "task_specific_params": null,
82
+ "problem_type": null,
83
+ "_name_or_path": "HuggingFaceTB/SmolLM2-360M-Instruct",
84
+ "_attn_implementation_autoset": true,
85
+ "transformers_version": "4.51.3",
86
+ "is_llama_config": true,
87
+ "model_type": "llama",
88
+ "rope_interleaved": false,
89
+ "transformers.js_config": {
90
+ "kv_cache_dtype": {
91
+ "q4f16": "float16",
92
+ "fp16": "float16"
93
+ }
94
+ }
95
+ },
96
+ {
97
+ "vocab_size": 49152,
98
+ "max_position_embeddings": 8192,
99
+ "hidden_size": 576,
100
+ "intermediate_size": 1536,
101
+ "num_hidden_layers": 30,
102
+ "num_attention_heads": 9,
103
+ "num_key_value_heads": 3,
104
+ "hidden_act": "silu",
105
+ "initializer_range": 0.041666666666666664,
106
+ "rms_norm_eps": 1e-05,
107
+ "pretraining_tp": 1,
108
+ "use_cache": true,
109
+ "rope_theta": 100000,
110
+ "rope_scaling": null,
111
+ "attention_bias": false,
112
+ "attention_dropout": 0.0,
113
+ "mlp_bias": false,
114
+ "head_dim": 64,
115
+ "return_dict": true,
116
+ "output_hidden_states": false,
117
+ "output_attentions": false,
118
+ "torchscript": false,
119
+ "torch_dtype": "float32",
120
+ "use_bfloat16": false,
121
+ "tf_legacy_loss": false,
122
+ "pruned_heads": {},
123
+ "tie_word_embeddings": true,
124
+ "chunk_size_feed_forward": 0,
125
+ "is_encoder_decoder": false,
126
+ "is_decoder": false,
127
+ "cross_attention_hidden_size": null,
128
+ "add_cross_attention": false,
129
+ "tie_encoder_decoder": false,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "do_sample": false,
133
+ "early_stopping": false,
134
+ "num_beams": 1,
135
+ "num_beam_groups": 1,
136
+ "diversity_penalty": 0.0,
137
+ "temperature": 1.0,
138
+ "top_k": 50,
139
+ "top_p": 1.0,
140
+ "typical_p": 1.0,
141
+ "repetition_penalty": 1.0,
142
+ "length_penalty": 1.0,
143
+ "no_repeat_ngram_size": 0,
144
+ "encoder_no_repeat_ngram_size": 0,
145
+ "bad_words_ids": null,
146
+ "num_return_sequences": 1,
147
+ "output_scores": false,
148
+ "return_dict_in_generate": false,
149
+ "forced_bos_token_id": null,
150
+ "forced_eos_token_id": null,
151
+ "remove_invalid_values": false,
152
+ "exponential_decay_length_penalty": null,
153
+ "suppress_tokens": null,
154
+ "begin_suppress_tokens": null,
155
+ "architectures": [
156
+ "LlamaForCausalLM"
157
+ ],
158
+ "finetuning_task": null,
159
+ "id2label": {
160
+ "0": "LABEL_0",
161
+ "1": "LABEL_1"
162
+ },
163
+ "label2id": {
164
+ "LABEL_0": 0,
165
+ "LABEL_1": 1
166
+ },
167
+ "tokenizer_class": null,
168
+ "prefix": null,
169
+ "bos_token_id": 1,
170
+ "pad_token_id": 2,
171
+ "eos_token_id": 2,
172
+ "sep_token_id": null,
173
+ "decoder_start_token_id": null,
174
+ "task_specific_params": null,
175
+ "problem_type": null,
176
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct",
177
+ "_attn_implementation_autoset": true,
178
+ "transformers_version": "4.51.3",
179
+ "is_llama_config": true,
180
+ "model_type": "llama",
181
+ "rope_interleaved": false,
182
+ "transformers.js_config": {
183
+ "kv_cache_dtype": {
184
+ "q4f16": "float16",
185
+ "fp16": "float16"
186
+ }
187
+ }
188
+ }
189
+ ],
190
+ "num_models": 2
191
+ }
checkpoint-2380/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2380/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
checkpoint-2380/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2380/tokenizer_config.json ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<repo_name>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "4": {
37
+ "content": "<reponame>",
38
+ "lstrip": false,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "5": {
45
+ "content": "<file_sep>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "6": {
53
+ "content": "<filename>",
54
+ "lstrip": false,
55
+ "normalized": false,
56
+ "rstrip": false,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "7": {
61
+ "content": "<gh_stars>",
62
+ "lstrip": false,
63
+ "normalized": false,
64
+ "rstrip": false,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "8": {
69
+ "content": "<issue_start>",
70
+ "lstrip": false,
71
+ "normalized": false,
72
+ "rstrip": false,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "9": {
77
+ "content": "<issue_comment>",
78
+ "lstrip": false,
79
+ "normalized": false,
80
+ "rstrip": false,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "10": {
85
+ "content": "<issue_closed>",
86
+ "lstrip": false,
87
+ "normalized": false,
88
+ "rstrip": false,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "11": {
93
+ "content": "<jupyter_start>",
94
+ "lstrip": false,
95
+ "normalized": false,
96
+ "rstrip": false,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "12": {
101
+ "content": "<jupyter_text>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "13": {
109
+ "content": "<jupyter_code>",
110
+ "lstrip": false,
111
+ "normalized": false,
112
+ "rstrip": false,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "14": {
117
+ "content": "<jupyter_output>",
118
+ "lstrip": false,
119
+ "normalized": false,
120
+ "rstrip": false,
121
+ "single_word": false,
122
+ "special": true
123
+ },
124
+ "15": {
125
+ "content": "<jupyter_script>",
126
+ "lstrip": false,
127
+ "normalized": false,
128
+ "rstrip": false,
129
+ "single_word": false,
130
+ "special": true
131
+ },
132
+ "16": {
133
+ "content": "<empty_output>",
134
+ "lstrip": false,
135
+ "normalized": false,
136
+ "rstrip": false,
137
+ "single_word": false,
138
+ "special": true
139
+ }
140
+ },
141
+ "additional_special_tokens": [
142
+ "<|im_start|>",
143
+ "<|im_end|>"
144
+ ],
145
+ "bos_token": "<|im_start|>",
146
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
147
+ "clean_up_tokenization_spaces": false,
148
+ "eos_token": "<|im_end|>",
149
+ "extra_special_tokens": {},
150
+ "model_max_length": 8192,
151
+ "pad_token": "<|im_end|>",
152
+ "tokenizer_class": "GPT2Tokenizer",
153
+ "unk_token": "<|endoftext|>",
154
+ "vocab_size": 49152
155
+ }
checkpoint-2380/trainer_state.json ADDED
@@ -0,0 +1,2216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 4.242532322781988,
6
+ "eval_steps": 500,
7
+ "global_step": 2380,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017833259028087384,
14
+ "grad_norm": 33.452999114990234,
15
+ "learning_rate": 3.2142857142857143e-06,
16
+ "loss": 8.9944,
17
+ "mean_token_accuracy": 0.32701094649964946,
18
+ "num_tokens": 2621440.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.03566651805617477,
23
+ "grad_norm": 2.596579074859619,
24
+ "learning_rate": 6.785714285714285e-06,
25
+ "loss": 4.9649,
26
+ "mean_token_accuracy": 0.5848119974136352,
27
+ "num_tokens": 5242880.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.05349977708426215,
32
+ "grad_norm": 1.8979898691177368,
33
+ "learning_rate": 1.0357142857142859e-05,
34
+ "loss": 4.1262,
35
+ "mean_token_accuracy": 0.5906830102205276,
36
+ "num_tokens": 7864320.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.07133303611234953,
41
+ "grad_norm": 1.1842058897018433,
42
+ "learning_rate": 1.392857142857143e-05,
43
+ "loss": 3.7327,
44
+ "mean_token_accuracy": 0.608327355235815,
45
+ "num_tokens": 10485760.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.08916629514043692,
50
+ "grad_norm": 0.922074556350708,
51
+ "learning_rate": 1.75e-05,
52
+ "loss": 3.4492,
53
+ "mean_token_accuracy": 0.6281662203371525,
54
+ "num_tokens": 13107200.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.1069995541685243,
59
+ "grad_norm": 0.7813541889190674,
60
+ "learning_rate": 2.107142857142857e-05,
61
+ "loss": 3.1416,
62
+ "mean_token_accuracy": 0.648241326212883,
63
+ "num_tokens": 15728640.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.12483281319661169,
68
+ "grad_norm": 1.1468125581741333,
69
+ "learning_rate": 2.4642857142857145e-05,
70
+ "loss": 2.8778,
71
+ "mean_token_accuracy": 0.6583510167896748,
72
+ "num_tokens": 18350080.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.14266607222469907,
77
+ "grad_norm": 1.0490431785583496,
78
+ "learning_rate": 2.8214285714285714e-05,
79
+ "loss": 2.7108,
80
+ "mean_token_accuracy": 0.663086223602295,
81
+ "num_tokens": 20971520.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.16049933125278645,
86
+ "grad_norm": 0.7813755869865417,
87
+ "learning_rate": 3.1785714285714284e-05,
88
+ "loss": 2.6039,
89
+ "mean_token_accuracy": 0.6650082431733608,
90
+ "num_tokens": 23592960.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.17833259028087384,
95
+ "grad_norm": 0.6500654220581055,
96
+ "learning_rate": 3.5357142857142864e-05,
97
+ "loss": 2.4513,
98
+ "mean_token_accuracy": 0.674950385093689,
99
+ "num_tokens": 26214400.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.19616584930896122,
104
+ "grad_norm": 0.81620192527771,
105
+ "learning_rate": 3.892857142857143e-05,
106
+ "loss": 2.3602,
107
+ "mean_token_accuracy": 0.6777540303766727,
108
+ "num_tokens": 28835840.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.2139991083370486,
113
+ "grad_norm": 0.6131238341331482,
114
+ "learning_rate": 4.25e-05,
115
+ "loss": 2.3545,
116
+ "mean_token_accuracy": 0.6758881151676178,
117
+ "num_tokens": 31457280.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.231832367365136,
122
+ "grad_norm": 0.9951689839363098,
123
+ "learning_rate": 4.607142857142857e-05,
124
+ "loss": 2.2624,
125
+ "mean_token_accuracy": 0.6806462205946445,
126
+ "num_tokens": 34078720.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.24966562639322337,
131
+ "grad_norm": 0.7796569466590881,
132
+ "learning_rate": 4.964285714285715e-05,
133
+ "loss": 2.1505,
134
+ "mean_token_accuracy": 0.6918394304811954,
135
+ "num_tokens": 36700160.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.26749888542131073,
140
+ "grad_norm": 0.8614610433578491,
141
+ "learning_rate": 4.9998587698517684e-05,
142
+ "loss": 2.1591,
143
+ "mean_token_accuracy": 0.68697941750288,
144
+ "num_tokens": 39321600.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.28533214444939814,
149
+ "grad_norm": 0.6178464889526367,
150
+ "learning_rate": 4.9993705873562665e-05,
151
+ "loss": 2.1048,
152
+ "mean_token_accuracy": 0.6911555036902428,
153
+ "num_tokens": 41943040.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.3031654034774855,
158
+ "grad_norm": 0.7255717515945435,
159
+ "learning_rate": 4.998533777009954e-05,
160
+ "loss": 2.0847,
161
+ "mean_token_accuracy": 0.6920825429260731,
162
+ "num_tokens": 44564480.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.3209986625055729,
167
+ "grad_norm": 0.7800336480140686,
168
+ "learning_rate": 4.9973484555363726e-05,
169
+ "loss": 2.049,
170
+ "mean_token_accuracy": 0.6982546880841255,
171
+ "num_tokens": 47185920.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.33883192153366026,
176
+ "grad_norm": 0.886745274066925,
177
+ "learning_rate": 4.995814788271582e-05,
178
+ "loss": 2.0301,
179
+ "mean_token_accuracy": 0.697366189956665,
180
+ "num_tokens": 49807360.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.3566651805617477,
185
+ "grad_norm": 0.6190801858901978,
186
+ "learning_rate": 4.993932989141088e-05,
187
+ "loss": 1.9408,
188
+ "mean_token_accuracy": 0.7095360569655895,
189
+ "num_tokens": 52428800.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.37449843958983503,
194
+ "grad_norm": 0.6333149671554565,
195
+ "learning_rate": 4.991703320630011e-05,
196
+ "loss": 1.9076,
197
+ "mean_token_accuracy": 0.7127362459897995,
198
+ "num_tokens": 55050240.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.39233169861792244,
203
+ "grad_norm": 0.9046627879142761,
204
+ "learning_rate": 4.989126093746468e-05,
205
+ "loss": 1.8913,
206
+ "mean_token_accuracy": 0.7148448958992958,
207
+ "num_tokens": 57671680.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.4101649576460098,
212
+ "grad_norm": 0.7778757214546204,
213
+ "learning_rate": 4.986201667978193e-05,
214
+ "loss": 1.9016,
215
+ "mean_token_accuracy": 0.7118340134620667,
216
+ "num_tokens": 60293120.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.4279982166740972,
221
+ "grad_norm": 0.6346383690834045,
222
+ "learning_rate": 4.9829304512423944e-05,
223
+ "loss": 1.898,
224
+ "mean_token_accuracy": 0.7123855009675026,
225
+ "num_tokens": 62914560.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.44583147570218457,
230
+ "grad_norm": 0.7308981418609619,
231
+ "learning_rate": 4.979312899828853e-05,
232
+ "loss": 1.8435,
233
+ "mean_token_accuracy": 0.7184088006615639,
234
+ "num_tokens": 65536000.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.463664734730272,
239
+ "grad_norm": 0.6174861788749695,
240
+ "learning_rate": 4.9753495183362796e-05,
241
+ "loss": 1.8242,
242
+ "mean_token_accuracy": 0.7238672457635402,
243
+ "num_tokens": 68157440.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.48149799375835933,
248
+ "grad_norm": 0.6507575511932373,
249
+ "learning_rate": 4.971040859601927e-05,
250
+ "loss": 1.791,
251
+ "mean_token_accuracy": 0.7251060992479325,
252
+ "num_tokens": 70778880.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.49933125278644674,
257
+ "grad_norm": 0.8092382550239563,
258
+ "learning_rate": 4.966387524624482e-05,
259
+ "loss": 1.7879,
260
+ "mean_token_accuracy": 0.7255102708935738,
261
+ "num_tokens": 73400320.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.5171645118145342,
266
+ "grad_norm": 0.5636172294616699,
267
+ "learning_rate": 4.9613901624802245e-05,
268
+ "loss": 1.7351,
269
+ "mean_token_accuracy": 0.7330663032829762,
270
+ "num_tokens": 76021760.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.5349977708426215,
275
+ "grad_norm": 0.655262291431427,
276
+ "learning_rate": 4.956049470232503e-05,
277
+ "loss": 1.7464,
278
+ "mean_token_accuracy": 0.7302443347871304,
279
+ "num_tokens": 78643200.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.5528310298707089,
284
+ "grad_norm": 0.9413332343101501,
285
+ "learning_rate": 4.9503661928344944e-05,
286
+ "loss": 1.7277,
287
+ "mean_token_accuracy": 0.730958030372858,
288
+ "num_tokens": 81264640.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.5706642888987963,
293
+ "grad_norm": 0.7439999580383301,
294
+ "learning_rate": 4.9443411230252956e-05,
295
+ "loss": 1.6947,
296
+ "mean_token_accuracy": 0.7346593342721462,
297
+ "num_tokens": 83886080.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.5884975479268837,
302
+ "grad_norm": 0.7665299773216248,
303
+ "learning_rate": 4.93797510121935e-05,
304
+ "loss": 1.7136,
305
+ "mean_token_accuracy": 0.7324949607253075,
306
+ "num_tokens": 86507520.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.606330806954971,
311
+ "grad_norm": 0.7366372346878052,
312
+ "learning_rate": 4.931269015389217e-05,
313
+ "loss": 1.676,
314
+ "mean_token_accuracy": 0.7371297933161258,
315
+ "num_tokens": 89128960.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.6241640659830584,
320
+ "grad_norm": 0.6571912169456482,
321
+ "learning_rate": 4.9242238009417175e-05,
322
+ "loss": 1.6252,
323
+ "mean_token_accuracy": 0.7434912383556366,
324
+ "num_tokens": 91750400.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.6419973250111458,
329
+ "grad_norm": 0.7496657967567444,
330
+ "learning_rate": 4.9168404405874505e-05,
331
+ "loss": 1.657,
332
+ "mean_token_accuracy": 0.737925547361374,
333
+ "num_tokens": 94371840.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.6598305840392331,
338
+ "grad_norm": 0.5275835394859314,
339
+ "learning_rate": 4.9091199642037244e-05,
340
+ "loss": 1.6723,
341
+ "mean_token_accuracy": 0.7358783438801766,
342
+ "num_tokens": 96993280.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.6776638430673205,
347
+ "grad_norm": 0.6143308281898499,
348
+ "learning_rate": 4.901063448690901e-05,
349
+ "loss": 1.6292,
350
+ "mean_token_accuracy": 0.7412570238113403,
351
+ "num_tokens": 99614720.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.6954971020954079,
356
+ "grad_norm": 0.860007643699646,
357
+ "learning_rate": 4.892672017822183e-05,
358
+ "loss": 1.6281,
359
+ "mean_token_accuracy": 0.7422710858285427,
360
+ "num_tokens": 102236160.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.7133303611234953,
365
+ "grad_norm": 0.5295814275741577,
366
+ "learning_rate": 4.8839468420868606e-05,
367
+ "loss": 1.6112,
368
+ "mean_token_accuracy": 0.7446457460522652,
369
+ "num_tokens": 104857600.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.7311636201515826,
374
+ "grad_norm": 0.7366165518760681,
375
+ "learning_rate": 4.8748891385270495e-05,
376
+ "loss": 1.6197,
377
+ "mean_token_accuracy": 0.7417318046092987,
378
+ "num_tokens": 107479040.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.7489968791796701,
383
+ "grad_norm": 0.7033573985099792,
384
+ "learning_rate": 4.865500170567926e-05,
385
+ "loss": 1.5763,
386
+ "mean_token_accuracy": 0.7474127545952797,
387
+ "num_tokens": 110100480.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.7668301382077575,
392
+ "grad_norm": 0.6527330875396729,
393
+ "learning_rate": 4.855781247841498e-05,
394
+ "loss": 1.5714,
395
+ "mean_token_accuracy": 0.7477886863052845,
396
+ "num_tokens": 112721920.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.7846633972358449,
401
+ "grad_norm": 0.7344551682472229,
402
+ "learning_rate": 4.845733726003928e-05,
403
+ "loss": 1.5682,
404
+ "mean_token_accuracy": 0.7475516766309738,
405
+ "num_tokens": 115343360.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.8024966562639322,
410
+ "grad_norm": 0.6840972304344177,
411
+ "learning_rate": 4.835359006546443e-05,
412
+ "loss": 1.5955,
413
+ "mean_token_accuracy": 0.7427977688610554,
414
+ "num_tokens": 117964800.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.8203299152920196,
419
+ "grad_norm": 0.635522723197937,
420
+ "learning_rate": 4.824658536599836e-05,
421
+ "loss": 1.5459,
422
+ "mean_token_accuracy": 0.7526196874678135,
423
+ "num_tokens": 120586240.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.838163174320107,
428
+ "grad_norm": 0.6874827146530151,
429
+ "learning_rate": 4.8136338087326216e-05,
430
+ "loss": 1.6108,
431
+ "mean_token_accuracy": 0.7413325920701027,
432
+ "num_tokens": 123207680.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.8559964333481944,
437
+ "grad_norm": 0.6571274995803833,
438
+ "learning_rate": 4.8022863607428356e-05,
439
+ "loss": 1.539,
440
+ "mean_token_accuracy": 0.7521994858980179,
441
+ "num_tokens": 125829120.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.8738296923762817,
446
+ "grad_norm": 0.8841336369514465,
447
+ "learning_rate": 4.790617775443537e-05,
448
+ "loss": 1.5486,
449
+ "mean_token_accuracy": 0.7492702730000019,
450
+ "num_tokens": 128450560.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.8916629514043691,
455
+ "grad_norm": 0.690990149974823,
456
+ "learning_rate": 4.7786296804420274e-05,
457
+ "loss": 1.5646,
458
+ "mean_token_accuracy": 0.7455468386411667,
459
+ "num_tokens": 131072000.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.9094962104324565,
464
+ "grad_norm": 0.7215222716331482,
465
+ "learning_rate": 4.766323747912818e-05,
466
+ "loss": 1.5085,
467
+ "mean_token_accuracy": 0.754824897646904,
468
+ "num_tokens": 133693440.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.927329469460544,
473
+ "grad_norm": 0.6024391651153564,
474
+ "learning_rate": 4.753701694364392e-05,
475
+ "loss": 1.4856,
476
+ "mean_token_accuracy": 0.7570274345576763,
477
+ "num_tokens": 136314880.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.9451627284886313,
482
+ "grad_norm": 0.6131455302238464,
483
+ "learning_rate": 4.740765280399767e-05,
484
+ "loss": 1.5433,
485
+ "mean_token_accuracy": 0.7495988786220551,
486
+ "num_tokens": 138936320.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.9629959875167187,
491
+ "grad_norm": 0.7875815629959106,
492
+ "learning_rate": 4.72751631047092e-05,
493
+ "loss": 1.4917,
494
+ "mean_token_accuracy": 0.7559954360127449,
495
+ "num_tokens": 141557760.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.9808292465448061,
500
+ "grad_norm": 0.7112794518470764,
501
+ "learning_rate": 4.71395663262709e-05,
502
+ "loss": 1.4784,
503
+ "mean_token_accuracy": 0.7572461232542992,
504
+ "num_tokens": 144179200.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.9986625055728935,
509
+ "grad_norm": 0.6807007789611816,
510
+ "learning_rate": 4.7000881382570014e-05,
511
+ "loss": 1.5159,
512
+ "mean_token_accuracy": 0.7523536741733551,
513
+ "num_tokens": 146800640.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 1.0,
518
+ "eval_loss": 1.494895339012146,
519
+ "eval_mean_token_accuracy": 0.7405715703964233,
520
+ "eval_num_tokens": 146980864.0,
521
+ "eval_runtime": 2.6056,
522
+ "eval_samples_per_second": 27.633,
523
+ "eval_steps_per_second": 1.919,
524
+ "step": 561
525
+ },
526
+ {
527
+ "epoch": 1.0160499331252786,
528
+ "grad_norm": 0.5480894446372986,
529
+ "learning_rate": 4.685912761825038e-05,
530
+ "loss": 1.4428,
531
+ "mean_token_accuracy": 0.7618689170250525,
532
+ "num_tokens": 149340160.0,
533
+ "step": 570
534
+ },
535
+ {
536
+ "epoch": 1.033883192153366,
537
+ "grad_norm": 0.6192435622215271,
538
+ "learning_rate": 4.671432480601416e-05,
539
+ "loss": 1.4711,
540
+ "mean_token_accuracy": 0.7556988887488842,
541
+ "num_tokens": 151961600.0,
542
+ "step": 580
543
+ },
544
+ {
545
+ "epoch": 1.0517164511814534,
546
+ "grad_norm": 0.5988848209381104,
547
+ "learning_rate": 4.656649314386378e-05,
548
+ "loss": 1.433,
549
+ "mean_token_accuracy": 0.7611474089324475,
550
+ "num_tokens": 154583040.0,
551
+ "step": 590
552
+ },
553
+ {
554
+ "epoch": 1.0695497102095408,
555
+ "grad_norm": 0.7051531672477722,
556
+ "learning_rate": 4.641565325228464e-05,
557
+ "loss": 1.415,
558
+ "mean_token_accuracy": 0.7643376767635346,
559
+ "num_tokens": 157204480.0,
560
+ "step": 600
561
+ },
562
+ {
563
+ "epoch": 1.0873829692376282,
564
+ "grad_norm": 0.6170579791069031,
565
+ "learning_rate": 4.6261826171368774e-05,
566
+ "loss": 1.4674,
567
+ "mean_token_accuracy": 0.7571297205984593,
568
+ "num_tokens": 159825920.0,
569
+ "step": 610
570
+ },
571
+ {
572
+ "epoch": 1.1052162282657156,
573
+ "grad_norm": 0.5691844820976257,
574
+ "learning_rate": 4.6105033357880104e-05,
575
+ "loss": 1.407,
576
+ "mean_token_accuracy": 0.7646078862249851,
577
+ "num_tokens": 162447360.0,
578
+ "step": 620
579
+ },
580
+ {
581
+ "epoch": 1.123049487293803,
582
+ "grad_norm": 0.5824171900749207,
583
+ "learning_rate": 4.594529668226146e-05,
584
+ "loss": 1.4185,
585
+ "mean_token_accuracy": 0.7614389970898628,
586
+ "num_tokens": 165068800.0,
587
+ "step": 630
588
+ },
589
+ {
590
+ "epoch": 1.1408827463218902,
591
+ "grad_norm": 0.5765758156776428,
592
+ "learning_rate": 4.578263842558402e-05,
593
+ "loss": 1.42,
594
+ "mean_token_accuracy": 0.7633033894002438,
595
+ "num_tokens": 167690240.0,
596
+ "step": 640
597
+ },
598
+ {
599
+ "epoch": 1.1587160053499777,
600
+ "grad_norm": 0.5261256694793701,
601
+ "learning_rate": 4.5617081276439355e-05,
602
+ "loss": 1.4384,
603
+ "mean_token_accuracy": 0.7604428693652153,
604
+ "num_tokens": 170311680.0,
605
+ "step": 650
606
+ },
607
+ {
608
+ "epoch": 1.176549264378065,
609
+ "grad_norm": 0.913208544254303,
610
+ "learning_rate": 4.544864832777469e-05,
611
+ "loss": 1.4775,
612
+ "mean_token_accuracy": 0.754877183586359,
613
+ "num_tokens": 172933120.0,
614
+ "step": 660
615
+ },
616
+ {
617
+ "epoch": 1.1943825234061525,
618
+ "grad_norm": 0.5799976587295532,
619
+ "learning_rate": 4.527736307367179e-05,
620
+ "loss": 1.4387,
621
+ "mean_token_accuracy": 0.7589173935353756,
622
+ "num_tokens": 175554560.0,
623
+ "step": 670
624
+ },
625
+ {
626
+ "epoch": 1.2122157824342399,
627
+ "grad_norm": 0.635047435760498,
628
+ "learning_rate": 4.510324940606979e-05,
629
+ "loss": 1.4439,
630
+ "mean_token_accuracy": 0.7597284145653248,
631
+ "num_tokens": 178176000.0,
632
+ "step": 680
633
+ },
634
+ {
635
+ "epoch": 1.2300490414623273,
636
+ "grad_norm": 0.5884873270988464,
637
+ "learning_rate": 4.4926331611432687e-05,
638
+ "loss": 1.3835,
639
+ "mean_token_accuracy": 0.7659787967801094,
640
+ "num_tokens": 180797440.0,
641
+ "step": 690
642
+ },
643
+ {
644
+ "epoch": 1.2478823004904147,
645
+ "grad_norm": 0.7232898473739624,
646
+ "learning_rate": 4.4746634367361634e-05,
647
+ "loss": 1.3629,
648
+ "mean_token_accuracy": 0.7691854700446129,
649
+ "num_tokens": 183418880.0,
650
+ "step": 700
651
+ },
652
+ {
653
+ "epoch": 1.265715559518502,
654
+ "grad_norm": 0.6569238305091858,
655
+ "learning_rate": 4.4564182739152805e-05,
656
+ "loss": 1.397,
657
+ "mean_token_accuracy": 0.7651830442249775,
658
+ "num_tokens": 186040320.0,
659
+ "step": 710
660
+ },
661
+ {
662
+ "epoch": 1.2835488185465893,
663
+ "grad_norm": 0.8630965352058411,
664
+ "learning_rate": 4.437900217630109e-05,
665
+ "loss": 1.4256,
666
+ "mean_token_accuracy": 0.7605882868170738,
667
+ "num_tokens": 188661760.0,
668
+ "step": 720
669
+ },
670
+ {
671
+ "epoch": 1.3013820775746767,
672
+ "grad_norm": 0.7600955367088318,
673
+ "learning_rate": 4.419111850895028e-05,
674
+ "loss": 1.3975,
675
+ "mean_token_accuracy": 0.7634797111153603,
676
+ "num_tokens": 191283200.0,
677
+ "step": 730
678
+ },
679
+ {
680
+ "epoch": 1.3192153366027641,
681
+ "grad_norm": 0.6687177419662476,
682
+ "learning_rate": 4.4000557944290086e-05,
683
+ "loss": 1.3616,
684
+ "mean_token_accuracy": 0.7703544825315476,
685
+ "num_tokens": 193904640.0,
686
+ "step": 740
687
+ },
688
+ {
689
+ "epoch": 1.3370485956308515,
690
+ "grad_norm": 0.719222903251648,
691
+ "learning_rate": 4.3807347062900624e-05,
692
+ "loss": 1.3788,
693
+ "mean_token_accuracy": 0.7648544363677502,
694
+ "num_tokens": 196526080.0,
695
+ "step": 750
696
+ },
697
+ {
698
+ "epoch": 1.354881854658939,
699
+ "grad_norm": 0.5986538529396057,
700
+ "learning_rate": 4.361151281504474e-05,
701
+ "loss": 1.364,
702
+ "mean_token_accuracy": 0.770479665696621,
703
+ "num_tokens": 199147520.0,
704
+ "step": 760
705
+ },
706
+ {
707
+ "epoch": 1.3727151136870264,
708
+ "grad_norm": 0.6472992897033691,
709
+ "learning_rate": 4.3413082516908893e-05,
710
+ "loss": 1.4045,
711
+ "mean_token_accuracy": 0.7644098080694676,
712
+ "num_tokens": 201768960.0,
713
+ "step": 770
714
+ },
715
+ {
716
+ "epoch": 1.3905483727151138,
717
+ "grad_norm": 0.44522207975387573,
718
+ "learning_rate": 4.321208384679285e-05,
719
+ "loss": 1.3314,
720
+ "mean_token_accuracy": 0.773588253557682,
721
+ "num_tokens": 204390400.0,
722
+ "step": 780
723
+ },
724
+ {
725
+ "epoch": 1.4083816317432012,
726
+ "grad_norm": 0.5019171833992004,
727
+ "learning_rate": 4.3008544841248986e-05,
728
+ "loss": 1.3591,
729
+ "mean_token_accuracy": 0.7695087313652038,
730
+ "num_tokens": 207011840.0,
731
+ "step": 790
732
+ },
733
+ {
734
+ "epoch": 1.4262148907712884,
735
+ "grad_norm": 0.7287567257881165,
736
+ "learning_rate": 4.280249389117157e-05,
737
+ "loss": 1.3773,
738
+ "mean_token_accuracy": 0.7675351880490779,
739
+ "num_tokens": 209633280.0,
740
+ "step": 800
741
+ },
742
+ {
743
+ "epoch": 1.4440481497993758,
744
+ "grad_norm": 1.0294618606567383,
745
+ "learning_rate": 4.2593959737836604e-05,
746
+ "loss": 1.3719,
747
+ "mean_token_accuracy": 0.7668005004525185,
748
+ "num_tokens": 212254720.0,
749
+ "step": 810
750
+ },
751
+ {
752
+ "epoch": 1.4618814088274632,
753
+ "grad_norm": 0.6181490421295166,
754
+ "learning_rate": 4.2382971468892806e-05,
755
+ "loss": 1.3168,
756
+ "mean_token_accuracy": 0.7758259050548076,
757
+ "num_tokens": 214876160.0,
758
+ "step": 820
759
+ },
760
+ {
761
+ "epoch": 1.4797146678555506,
762
+ "grad_norm": 0.5945810079574585,
763
+ "learning_rate": 4.216955851430432e-05,
764
+ "loss": 1.3536,
765
+ "mean_token_accuracy": 0.7710937492549419,
766
+ "num_tokens": 217497600.0,
767
+ "step": 830
768
+ },
769
+ {
770
+ "epoch": 1.497547926883638,
771
+ "grad_norm": 0.6803934574127197,
772
+ "learning_rate": 4.195375064224562e-05,
773
+ "loss": 1.324,
774
+ "mean_token_accuracy": 0.7753530308604241,
775
+ "num_tokens": 220119040.0,
776
+ "step": 840
777
+ },
778
+ {
779
+ "epoch": 1.5153811859117252,
780
+ "grad_norm": 0.5748502612113953,
781
+ "learning_rate": 4.1735577954949225e-05,
782
+ "loss": 1.3057,
783
+ "mean_token_accuracy": 0.7755480580031872,
784
+ "num_tokens": 222740480.0,
785
+ "step": 850
786
+ },
787
+ {
788
+ "epoch": 1.5332144449398126,
789
+ "grad_norm": 0.5675086379051208,
790
+ "learning_rate": 4.151507088450692e-05,
791
+ "loss": 1.302,
792
+ "mean_token_accuracy": 0.7765964694321156,
793
+ "num_tokens": 225361920.0,
794
+ "step": 860
795
+ },
796
+ {
797
+ "epoch": 1.5510477039679,
798
+ "grad_norm": 0.6266924142837524,
799
+ "learning_rate": 4.129226018862484e-05,
800
+ "loss": 1.353,
801
+ "mean_token_accuracy": 0.770804837346077,
802
+ "num_tokens": 227983360.0,
803
+ "step": 870
804
+ },
805
+ {
806
+ "epoch": 1.5688809629959874,
807
+ "grad_norm": 0.551527202129364,
808
+ "learning_rate": 4.106717694633321e-05,
809
+ "loss": 1.3477,
810
+ "mean_token_accuracy": 0.77206010222435,
811
+ "num_tokens": 230604800.0,
812
+ "step": 880
813
+ },
814
+ {
815
+ "epoch": 1.5867142220240749,
816
+ "grad_norm": 0.6628434658050537,
817
+ "learning_rate": 4.0839852553651265e-05,
818
+ "loss": 1.3244,
819
+ "mean_token_accuracy": 0.773486353456974,
820
+ "num_tokens": 233226240.0,
821
+ "step": 890
822
+ },
823
+ {
824
+ "epoch": 1.6045474810521623,
825
+ "grad_norm": 0.5956223011016846,
826
+ "learning_rate": 4.0610318719207906e-05,
827
+ "loss": 1.3298,
828
+ "mean_token_accuracy": 0.7733359761536122,
829
+ "num_tokens": 235847680.0,
830
+ "step": 900
831
+ },
832
+ {
833
+ "epoch": 1.6223807400802497,
834
+ "grad_norm": 0.6573621034622192,
835
+ "learning_rate": 4.037860745981881e-05,
836
+ "loss": 1.283,
837
+ "mean_token_accuracy": 0.7787142746150494,
838
+ "num_tokens": 238469120.0,
839
+ "step": 910
840
+ },
841
+ {
842
+ "epoch": 1.640213999108337,
843
+ "grad_norm": 0.607742965221405,
844
+ "learning_rate": 4.01447510960205e-05,
845
+ "loss": 1.2955,
846
+ "mean_token_accuracy": 0.7780937045812607,
847
+ "num_tokens": 241090560.0,
848
+ "step": 920
849
+ },
850
+ {
851
+ "epoch": 1.6580472581364245,
852
+ "grad_norm": 0.6278334856033325,
853
+ "learning_rate": 3.99087822475621e-05,
854
+ "loss": 1.3215,
855
+ "mean_token_accuracy": 0.7741382211446762,
856
+ "num_tokens": 243712000.0,
857
+ "step": 930
858
+ },
859
+ {
860
+ "epoch": 1.675880517164512,
861
+ "grad_norm": 0.5402811169624329,
862
+ "learning_rate": 3.96707338288553e-05,
863
+ "loss": 1.3113,
864
+ "mean_token_accuracy": 0.7751454092562199,
865
+ "num_tokens": 246333440.0,
866
+ "step": 940
867
+ },
868
+ {
869
+ "epoch": 1.6937137761925993,
870
+ "grad_norm": 0.6165571212768555,
871
+ "learning_rate": 3.9430639044383274e-05,
872
+ "loss": 1.3093,
873
+ "mean_token_accuracy": 0.7768628627061844,
874
+ "num_tokens": 248954880.0,
875
+ "step": 950
876
+ },
877
+ {
878
+ "epoch": 1.7115470352206867,
879
+ "grad_norm": 0.6086046695709229,
880
+ "learning_rate": 3.9188531384069096e-05,
881
+ "loss": 1.3125,
882
+ "mean_token_accuracy": 0.7762621335685254,
883
+ "num_tokens": 251576320.0,
884
+ "step": 960
885
+ },
886
+ {
887
+ "epoch": 1.729380294248774,
888
+ "grad_norm": 0.682686984539032,
889
+ "learning_rate": 3.8944444618604356e-05,
890
+ "loss": 1.2821,
891
+ "mean_token_accuracy": 0.7795726217329502,
892
+ "num_tokens": 254197760.0,
893
+ "step": 970
894
+ },
895
+ {
896
+ "epoch": 1.7472135532768613,
897
+ "grad_norm": 0.5615099668502808,
898
+ "learning_rate": 3.869841279473862e-05,
899
+ "loss": 1.3095,
900
+ "mean_token_accuracy": 0.7766918800771236,
901
+ "num_tokens": 256819200.0,
902
+ "step": 980
903
+ },
904
+ {
905
+ "epoch": 1.7650468123049488,
906
+ "grad_norm": 0.5341463088989258,
907
+ "learning_rate": 3.845047023053037e-05,
908
+ "loss": 1.2864,
909
+ "mean_token_accuracy": 0.779373399913311,
910
+ "num_tokens": 259440640.0,
911
+ "step": 990
912
+ },
913
+ {
914
+ "epoch": 1.7828800713330362,
915
+ "grad_norm": 0.5641157627105713,
916
+ "learning_rate": 3.82006515105601e-05,
917
+ "loss": 1.2857,
918
+ "mean_token_accuracy": 0.7798695519566536,
919
+ "num_tokens": 262062080.0,
920
+ "step": 1000
921
+ },
922
+ {
923
+ "epoch": 1.8007133303611234,
924
+ "grad_norm": 0.5196830630302429,
925
+ "learning_rate": 3.794899148110628e-05,
926
+ "loss": 1.2949,
927
+ "mean_token_accuracy": 0.7788909852504731,
928
+ "num_tokens": 264683520.0,
929
+ "step": 1010
930
+ },
931
+ {
932
+ "epoch": 1.8185465893892108,
933
+ "grad_norm": 0.4990008771419525,
934
+ "learning_rate": 3.7695525245284716e-05,
935
+ "loss": 1.254,
936
+ "mean_token_accuracy": 0.7839147254824639,
937
+ "num_tokens": 267304960.0,
938
+ "step": 1020
939
+ },
940
+ {
941
+ "epoch": 1.8363798484172982,
942
+ "grad_norm": 0.5643882751464844,
943
+ "learning_rate": 3.7440288158152187e-05,
944
+ "loss": 1.2446,
945
+ "mean_token_accuracy": 0.785726061463356,
946
+ "num_tokens": 269926400.0,
947
+ "step": 1030
948
+ },
949
+ {
950
+ "epoch": 1.8542131074453856,
951
+ "grad_norm": 0.8107313513755798,
952
+ "learning_rate": 3.7183315821774914e-05,
953
+ "loss": 1.2974,
954
+ "mean_token_accuracy": 0.7775464087724686,
955
+ "num_tokens": 272547840.0,
956
+ "step": 1040
957
+ },
958
+ {
959
+ "epoch": 1.872046366473473,
960
+ "grad_norm": 0.5176706314086914,
961
+ "learning_rate": 3.692464408026253e-05,
962
+ "loss": 1.2916,
963
+ "mean_token_accuracy": 0.7772151328623295,
964
+ "num_tokens": 275169280.0,
965
+ "step": 1050
966
+ },
967
+ {
968
+ "epoch": 1.8898796255015604,
969
+ "grad_norm": 0.5898165106773376,
970
+ "learning_rate": 3.666430901476833e-05,
971
+ "loss": 1.2716,
972
+ "mean_token_accuracy": 0.7806546173989772,
973
+ "num_tokens": 277790720.0,
974
+ "step": 1060
975
+ },
976
+ {
977
+ "epoch": 1.9077128845296478,
978
+ "grad_norm": 0.5368467569351196,
979
+ "learning_rate": 3.640234693845645e-05,
980
+ "loss": 1.2491,
981
+ "mean_token_accuracy": 0.7844746112823486,
982
+ "num_tokens": 280412160.0,
983
+ "step": 1070
984
+ },
985
+ {
986
+ "epoch": 1.9255461435577352,
987
+ "grad_norm": 0.5310569405555725,
988
+ "learning_rate": 3.613879439143667e-05,
989
+ "loss": 1.2825,
990
+ "mean_token_accuracy": 0.7795199528336525,
991
+ "num_tokens": 283033600.0,
992
+ "step": 1080
993
+ },
994
+ {
995
+ "epoch": 1.9433794025858226,
996
+ "grad_norm": 0.6114910840988159,
997
+ "learning_rate": 3.58736881356676e-05,
998
+ "loss": 1.2595,
999
+ "mean_token_accuracy": 0.7831220239400863,
1000
+ "num_tokens": 285655040.0,
1001
+ "step": 1090
1002
+ },
1003
+ {
1004
+ "epoch": 1.96121266161391,
1005
+ "grad_norm": 0.5786811113357544,
1006
+ "learning_rate": 3.5607065149828843e-05,
1007
+ "loss": 1.2694,
1008
+ "mean_token_accuracy": 0.7822808526456356,
1009
+ "num_tokens": 288276480.0,
1010
+ "step": 1100
1011
+ },
1012
+ {
1013
+ "epoch": 1.9790459206419975,
1014
+ "grad_norm": 0.6175279021263123,
1015
+ "learning_rate": 3.533896262416302e-05,
1016
+ "loss": 1.2712,
1017
+ "mean_token_accuracy": 0.7803981445729733,
1018
+ "num_tokens": 290897920.0,
1019
+ "step": 1110
1020
+ },
1021
+ {
1022
+ "epoch": 1.9968791796700847,
1023
+ "grad_norm": 0.568803608417511,
1024
+ "learning_rate": 3.506941795528821e-05,
1025
+ "loss": 1.2576,
1026
+ "mean_token_accuracy": 0.7821117818355561,
1027
+ "num_tokens": 293519360.0,
1028
+ "step": 1120
1029
+ },
1030
+ {
1031
+ "epoch": 2.0,
1032
+ "eval_loss": 1.2395988702774048,
1033
+ "eval_mean_token_accuracy": 0.7735649704933166,
1034
+ "eval_num_tokens": 293961728.0,
1035
+ "eval_runtime": 2.8804,
1036
+ "eval_samples_per_second": 24.997,
1037
+ "eval_steps_per_second": 1.736,
1038
+ "step": 1122
1039
+ },
1040
+ {
1041
+ "epoch": 2.0142666072224698,
1042
+ "grad_norm": 0.6652095913887024,
1043
+ "learning_rate": 3.479846874098167e-05,
1044
+ "loss": 1.2741,
1045
+ "mean_token_accuracy": 0.7794669828353784,
1046
+ "num_tokens": 296058880.0,
1047
+ "step": 1130
1048
+ },
1049
+ {
1050
+ "epoch": 2.032099866250557,
1051
+ "grad_norm": 0.7581707835197449,
1052
+ "learning_rate": 3.452615277493544e-05,
1053
+ "loss": 1.199,
1054
+ "mean_token_accuracy": 0.7900658771395683,
1055
+ "num_tokens": 298680320.0,
1056
+ "step": 1140
1057
+ },
1058
+ {
1059
+ "epoch": 2.0499331252786446,
1060
+ "grad_norm": 0.5383137464523315,
1061
+ "learning_rate": 3.425250804148464e-05,
1062
+ "loss": 1.2304,
1063
+ "mean_token_accuracy": 0.7861943542957306,
1064
+ "num_tokens": 301301760.0,
1065
+ "step": 1150
1066
+ },
1067
+ {
1068
+ "epoch": 2.067766384306732,
1069
+ "grad_norm": 0.5952672958374023,
1070
+ "learning_rate": 3.3977572710309166e-05,
1071
+ "loss": 1.2445,
1072
+ "mean_token_accuracy": 0.7836834371089936,
1073
+ "num_tokens": 303923200.0,
1074
+ "step": 1160
1075
+ },
1076
+ {
1077
+ "epoch": 2.0855996433348194,
1078
+ "grad_norm": 0.646511435508728,
1079
+ "learning_rate": 3.3701385131109616e-05,
1080
+ "loss": 1.2019,
1081
+ "mean_token_accuracy": 0.7892128698527813,
1082
+ "num_tokens": 306544640.0,
1083
+ "step": 1170
1084
+ },
1085
+ {
1086
+ "epoch": 2.103432902362907,
1087
+ "grad_norm": 0.514665961265564,
1088
+ "learning_rate": 3.342398382825791e-05,
1089
+ "loss": 1.2614,
1090
+ "mean_token_accuracy": 0.7812080189585686,
1091
+ "num_tokens": 309166080.0,
1092
+ "step": 1180
1093
+ },
1094
+ {
1095
+ "epoch": 2.121266161390994,
1096
+ "grad_norm": 0.5490918159484863,
1097
+ "learning_rate": 3.31454074954238e-05,
1098
+ "loss": 1.2108,
1099
+ "mean_token_accuracy": 0.7876022845506668,
1100
+ "num_tokens": 311787520.0,
1101
+ "step": 1190
1102
+ },
1103
+ {
1104
+ "epoch": 2.1390994204190816,
1105
+ "grad_norm": 0.6549789309501648,
1106
+ "learning_rate": 3.286569499017752e-05,
1107
+ "loss": 1.2024,
1108
+ "mean_token_accuracy": 0.7898712314665317,
1109
+ "num_tokens": 314408960.0,
1110
+ "step": 1200
1111
+ },
1112
+ {
1113
+ "epoch": 2.156932679447169,
1114
+ "grad_norm": 0.5743107199668884,
1115
+ "learning_rate": 3.2584885328569785e-05,
1116
+ "loss": 1.1835,
1117
+ "mean_token_accuracy": 0.7928275369107723,
1118
+ "num_tokens": 317030400.0,
1119
+ "step": 1210
1120
+ },
1121
+ {
1122
+ "epoch": 2.1747659384752565,
1123
+ "grad_norm": 0.6201761960983276,
1124
+ "learning_rate": 3.2303017679689525e-05,
1125
+ "loss": 1.1939,
1126
+ "mean_token_accuracy": 0.7908047571778297,
1127
+ "num_tokens": 319651840.0,
1128
+ "step": 1220
1129
+ },
1130
+ {
1131
+ "epoch": 2.192599197503344,
1132
+ "grad_norm": 0.550830602645874,
1133
+ "learning_rate": 3.2020131360200354e-05,
1134
+ "loss": 1.1864,
1135
+ "mean_token_accuracy": 0.7930233284831048,
1136
+ "num_tokens": 322273280.0,
1137
+ "step": 1230
1138
+ },
1139
+ {
1140
+ "epoch": 2.2104324565314313,
1141
+ "grad_norm": 0.5972590446472168,
1142
+ "learning_rate": 3.173626582885645e-05,
1143
+ "loss": 1.2177,
1144
+ "mean_token_accuracy": 0.7885610058903694,
1145
+ "num_tokens": 324894720.0,
1146
+ "step": 1240
1147
+ },
1148
+ {
1149
+ "epoch": 2.2282657155595187,
1150
+ "grad_norm": 0.5069220066070557,
1151
+ "learning_rate": 3.145146068099858e-05,
1152
+ "loss": 1.2279,
1153
+ "mean_token_accuracy": 0.7862870991230011,
1154
+ "num_tokens": 327516160.0,
1155
+ "step": 1250
1156
+ },
1157
+ {
1158
+ "epoch": 2.246098974587606,
1159
+ "grad_norm": 0.5100940465927124,
1160
+ "learning_rate": 3.116575564303109e-05,
1161
+ "loss": 1.1777,
1162
+ "mean_token_accuracy": 0.7929233327507973,
1163
+ "num_tokens": 330137600.0,
1164
+ "step": 1260
1165
+ },
1166
+ {
1167
+ "epoch": 2.2639322336156935,
1168
+ "grad_norm": 0.6175646781921387,
1169
+ "learning_rate": 3.087919056688067e-05,
1170
+ "loss": 1.173,
1171
+ "mean_token_accuracy": 0.7950430497527122,
1172
+ "num_tokens": 332759040.0,
1173
+ "step": 1270
1174
+ },
1175
+ {
1176
+ "epoch": 2.2817654926437805,
1177
+ "grad_norm": 0.5933928489685059,
1178
+ "learning_rate": 3.059180542443746e-05,
1179
+ "loss": 1.2057,
1180
+ "mean_token_accuracy": 0.7882316343486309,
1181
+ "num_tokens": 335380480.0,
1182
+ "step": 1280
1183
+ },
1184
+ {
1185
+ "epoch": 2.299598751671868,
1186
+ "grad_norm": 0.4882568120956421,
1187
+ "learning_rate": 3.0303640301979635e-05,
1188
+ "loss": 1.1982,
1189
+ "mean_token_accuracy": 0.7903318881988526,
1190
+ "num_tokens": 338001920.0,
1191
+ "step": 1290
1192
+ },
1193
+ {
1194
+ "epoch": 2.3174320106999553,
1195
+ "grad_norm": 0.5135630369186401,
1196
+ "learning_rate": 3.0014735394581823e-05,
1197
+ "loss": 1.2045,
1198
+ "mean_token_accuracy": 0.7893250770866871,
1199
+ "num_tokens": 340623360.0,
1200
+ "step": 1300
1201
+ },
1202
+ {
1203
+ "epoch": 2.3352652697280427,
1204
+ "grad_norm": 0.5496548414230347,
1205
+ "learning_rate": 2.972513100050851e-05,
1206
+ "loss": 1.1883,
1207
+ "mean_token_accuracy": 0.791755847632885,
1208
+ "num_tokens": 343244800.0,
1209
+ "step": 1310
1210
+ },
1211
+ {
1212
+ "epoch": 2.35309852875613,
1213
+ "grad_norm": 0.5601444840431213,
1214
+ "learning_rate": 2.943486751559295e-05,
1215
+ "loss": 1.1263,
1216
+ "mean_token_accuracy": 0.7998171858489513,
1217
+ "num_tokens": 345866240.0,
1218
+ "step": 1320
1219
+ },
1220
+ {
1221
+ "epoch": 2.3709317877842175,
1222
+ "grad_norm": 0.5243608355522156,
1223
+ "learning_rate": 2.9143985427602516e-05,
1224
+ "loss": 1.2274,
1225
+ "mean_token_accuracy": 0.7866500526666641,
1226
+ "num_tokens": 348487680.0,
1227
+ "step": 1330
1228
+ },
1229
+ {
1230
+ "epoch": 2.388765046812305,
1231
+ "grad_norm": 0.573567807674408,
1232
+ "learning_rate": 2.88525253105912e-05,
1233
+ "loss": 1.1977,
1234
+ "mean_token_accuracy": 0.7901960171759128,
1235
+ "num_tokens": 351109120.0,
1236
+ "step": 1340
1237
+ },
1238
+ {
1239
+ "epoch": 2.4065983058403924,
1240
+ "grad_norm": 0.4823901951313019,
1241
+ "learning_rate": 2.856052781924012e-05,
1242
+ "loss": 1.1358,
1243
+ "mean_token_accuracy": 0.7998057343065739,
1244
+ "num_tokens": 353730560.0,
1245
+ "step": 1350
1246
+ },
1247
+ {
1248
+ "epoch": 2.4244315648684798,
1249
+ "grad_norm": 0.515941858291626,
1250
+ "learning_rate": 2.8268033683186697e-05,
1251
+ "loss": 1.1881,
1252
+ "mean_token_accuracy": 0.7909879550337792,
1253
+ "num_tokens": 356352000.0,
1254
+ "step": 1360
1255
+ },
1256
+ {
1257
+ "epoch": 2.442264823896567,
1258
+ "grad_norm": 0.5508431196212769,
1259
+ "learning_rate": 2.7975083701343497e-05,
1260
+ "loss": 1.171,
1261
+ "mean_token_accuracy": 0.7945770494639873,
1262
+ "num_tokens": 358973440.0,
1263
+ "step": 1370
1264
+ },
1265
+ {
1266
+ "epoch": 2.4600980829246546,
1267
+ "grad_norm": 0.5473787188529968,
1268
+ "learning_rate": 2.7681718736207298e-05,
1269
+ "loss": 1.1309,
1270
+ "mean_token_accuracy": 0.7995454497635365,
1271
+ "num_tokens": 361594880.0,
1272
+ "step": 1380
1273
+ },
1274
+ {
1275
+ "epoch": 2.477931341952742,
1276
+ "grad_norm": 0.6680080890655518,
1277
+ "learning_rate": 2.738797970815933e-05,
1278
+ "loss": 1.1848,
1279
+ "mean_token_accuracy": 0.7917638584971428,
1280
+ "num_tokens": 364216320.0,
1281
+ "step": 1390
1282
+ },
1283
+ {
1284
+ "epoch": 2.4957646009808294,
1285
+ "grad_norm": 0.49958300590515137,
1286
+ "learning_rate": 2.7093907589757468e-05,
1287
+ "loss": 1.1922,
1288
+ "mean_token_accuracy": 0.7899307683110237,
1289
+ "num_tokens": 366837760.0,
1290
+ "step": 1400
1291
+ },
1292
+ {
1293
+ "epoch": 2.5135978600089164,
1294
+ "grad_norm": 0.4601416289806366,
1295
+ "learning_rate": 2.6799543400021132e-05,
1296
+ "loss": 1.1448,
1297
+ "mean_token_accuracy": 0.7981104202568531,
1298
+ "num_tokens": 369459200.0,
1299
+ "step": 1410
1300
+ },
1301
+ {
1302
+ "epoch": 2.531431119037004,
1303
+ "grad_norm": 0.5619862079620361,
1304
+ "learning_rate": 2.650492819870965e-05,
1305
+ "loss": 1.1371,
1306
+ "mean_token_accuracy": 0.7994878143072128,
1307
+ "num_tokens": 372080640.0,
1308
+ "step": 1420
1309
+ },
1310
+ {
1311
+ "epoch": 2.549264378065091,
1312
+ "grad_norm": 0.6030861139297485,
1313
+ "learning_rate": 2.621010308059504e-05,
1314
+ "loss": 1.1695,
1315
+ "mean_token_accuracy": 0.7944411806762218,
1316
+ "num_tokens": 374702080.0,
1317
+ "step": 1430
1318
+ },
1319
+ {
1320
+ "epoch": 2.5670976370931786,
1321
+ "grad_norm": 0.5237954258918762,
1322
+ "learning_rate": 2.5915109169729814e-05,
1323
+ "loss": 1.1372,
1324
+ "mean_token_accuracy": 0.7982218623161316,
1325
+ "num_tokens": 377323520.0,
1326
+ "step": 1440
1327
+ },
1328
+ {
1329
+ "epoch": 2.584930896121266,
1330
+ "grad_norm": 0.5144650936126709,
1331
+ "learning_rate": 2.5619987613710756e-05,
1332
+ "loss": 1.1896,
1333
+ "mean_token_accuracy": 0.7915455557405948,
1334
+ "num_tokens": 379944960.0,
1335
+ "step": 1450
1336
+ },
1337
+ {
1338
+ "epoch": 2.6027641551493534,
1339
+ "grad_norm": 0.5404797196388245,
1340
+ "learning_rate": 2.5324779577939394e-05,
1341
+ "loss": 1.1364,
1342
+ "mean_token_accuracy": 0.7987703062593937,
1343
+ "num_tokens": 382566400.0,
1344
+ "step": 1460
1345
+ },
1346
+ {
1347
+ "epoch": 2.620597414177441,
1348
+ "grad_norm": 0.4770708382129669,
1349
+ "learning_rate": 2.5029526239880015e-05,
1350
+ "loss": 1.1541,
1351
+ "mean_token_accuracy": 0.7973127588629723,
1352
+ "num_tokens": 385187840.0,
1353
+ "step": 1470
1354
+ },
1355
+ {
1356
+ "epoch": 2.6384306732055283,
1357
+ "grad_norm": 0.5579174757003784,
1358
+ "learning_rate": 2.473426878331591e-05,
1359
+ "loss": 1.1695,
1360
+ "mean_token_accuracy": 0.7945644512772561,
1361
+ "num_tokens": 387809280.0,
1362
+ "step": 1480
1363
+ },
1364
+ {
1365
+ "epoch": 2.6562639322336157,
1366
+ "grad_norm": 0.5287127494812012,
1367
+ "learning_rate": 2.443904839260488e-05,
1368
+ "loss": 1.1615,
1369
+ "mean_token_accuracy": 0.7949182473123073,
1370
+ "num_tokens": 390430720.0,
1371
+ "step": 1490
1372
+ },
1373
+ {
1374
+ "epoch": 2.674097191261703,
1375
+ "grad_norm": 0.5106226801872253,
1376
+ "learning_rate": 2.4143906246934537e-05,
1377
+ "loss": 1.1201,
1378
+ "mean_token_accuracy": 0.8007564425468445,
1379
+ "num_tokens": 393052160.0,
1380
+ "step": 1500
1381
+ },
1382
+ {
1383
+ "epoch": 2.6919304502897905,
1384
+ "grad_norm": 0.48907479643821716,
1385
+ "learning_rate": 2.384888351457838e-05,
1386
+ "loss": 1.1908,
1387
+ "mean_token_accuracy": 0.7911803111433983,
1388
+ "num_tokens": 395673600.0,
1389
+ "step": 1510
1390
+ },
1391
+ {
1392
+ "epoch": 2.709763709317878,
1393
+ "grad_norm": 0.5045855045318604,
1394
+ "learning_rate": 2.3554021347153403e-05,
1395
+ "loss": 1.1483,
1396
+ "mean_token_accuracy": 0.7968017213046551,
1397
+ "num_tokens": 398295040.0,
1398
+ "step": 1520
1399
+ },
1400
+ {
1401
+ "epoch": 2.7275969683459653,
1402
+ "grad_norm": 0.505662739276886,
1403
+ "learning_rate": 2.3259360873879992e-05,
1404
+ "loss": 1.1749,
1405
+ "mean_token_accuracy": 0.7930756144225597,
1406
+ "num_tokens": 400916480.0,
1407
+ "step": 1530
1408
+ },
1409
+ {
1410
+ "epoch": 2.7454302273740527,
1411
+ "grad_norm": 0.5283376574516296,
1412
+ "learning_rate": 2.2964943195844966e-05,
1413
+ "loss": 1.173,
1414
+ "mean_token_accuracy": 0.7921638391911984,
1415
+ "num_tokens": 403537920.0,
1416
+ "step": 1540
1417
+ },
1418
+ {
1419
+ "epoch": 2.76326348640214,
1420
+ "grad_norm": 0.5634790062904358,
1421
+ "learning_rate": 2.2670809380268548e-05,
1422
+ "loss": 1.1779,
1423
+ "mean_token_accuracy": 0.7936561174690724,
1424
+ "num_tokens": 406159360.0,
1425
+ "step": 1550
1426
+ },
1427
+ {
1428
+ "epoch": 2.7810967454302276,
1429
+ "grad_norm": 0.48228803277015686,
1430
+ "learning_rate": 2.237700045477607e-05,
1431
+ "loss": 1.14,
1432
+ "mean_token_accuracy": 0.796902097016573,
1433
+ "num_tokens": 408780800.0,
1434
+ "step": 1560
1435
+ },
1436
+ {
1437
+ "epoch": 2.798930004458315,
1438
+ "grad_norm": 0.4753395617008209,
1439
+ "learning_rate": 2.2083557401675115e-05,
1440
+ "loss": 1.1294,
1441
+ "mean_token_accuracy": 0.7987142026424408,
1442
+ "num_tokens": 411402240.0,
1443
+ "step": 1570
1444
+ },
1445
+ {
1446
+ "epoch": 2.8167632634864024,
1447
+ "grad_norm": 0.5348941683769226,
1448
+ "learning_rate": 2.1790521152239137e-05,
1449
+ "loss": 1.1581,
1450
+ "mean_token_accuracy": 0.7944865994155407,
1451
+ "num_tokens": 414023680.0,
1452
+ "step": 1580
1453
+ },
1454
+ {
1455
+ "epoch": 2.83459652251449,
1456
+ "grad_norm": 0.5610408782958984,
1457
+ "learning_rate": 2.1497932580998053e-05,
1458
+ "loss": 1.1609,
1459
+ "mean_token_accuracy": 0.795001070201397,
1460
+ "num_tokens": 416645120.0,
1461
+ "step": 1590
1462
+ },
1463
+ {
1464
+ "epoch": 2.8524297815425768,
1465
+ "grad_norm": 0.5431333184242249,
1466
+ "learning_rate": 2.1205832500036838e-05,
1467
+ "loss": 1.1405,
1468
+ "mean_token_accuracy": 0.7983539171516896,
1469
+ "num_tokens": 419266560.0,
1470
+ "step": 1600
1471
+ },
1472
+ {
1473
+ "epoch": 2.870263040570664,
1474
+ "grad_norm": 0.5650882720947266,
1475
+ "learning_rate": 2.0914261653302747e-05,
1476
+ "loss": 1.17,
1477
+ "mean_token_accuracy": 0.7925355695188046,
1478
+ "num_tokens": 421888000.0,
1479
+ "step": 1610
1480
+ },
1481
+ {
1482
+ "epoch": 2.8880962995987516,
1483
+ "grad_norm": 0.5435991287231445,
1484
+ "learning_rate": 2.062326071092218e-05,
1485
+ "loss": 1.1562,
1486
+ "mean_token_accuracy": 0.7965150982141495,
1487
+ "num_tokens": 424509440.0,
1488
+ "step": 1620
1489
+ },
1490
+ {
1491
+ "epoch": 2.905929558626839,
1492
+ "grad_norm": 0.4466145932674408,
1493
+ "learning_rate": 2.0332870263527675e-05,
1494
+ "loss": 1.1199,
1495
+ "mean_token_accuracy": 0.8016460791230202,
1496
+ "num_tokens": 427130880.0,
1497
+ "step": 1630
1498
+ },
1499
+ {
1500
+ "epoch": 2.9237628176549264,
1501
+ "grad_norm": 0.5083450675010681,
1502
+ "learning_rate": 2.0043130816596136e-05,
1503
+ "loss": 1.1467,
1504
+ "mean_token_accuracy": 0.7959990955889225,
1505
+ "num_tokens": 429752320.0,
1506
+ "step": 1640
1507
+ },
1508
+ {
1509
+ "epoch": 2.941596076683014,
1510
+ "grad_norm": 0.49085912108421326,
1511
+ "learning_rate": 1.9754082784798843e-05,
1512
+ "loss": 1.1565,
1513
+ "mean_token_accuracy": 0.7959273479878902,
1514
+ "num_tokens": 432373760.0,
1515
+ "step": 1650
1516
+ },
1517
+ {
1518
+ "epoch": 2.9594293357111012,
1519
+ "grad_norm": 0.48444733023643494,
1520
+ "learning_rate": 1.9465766486364143e-05,
1521
+ "loss": 1.1342,
1522
+ "mean_token_accuracy": 0.797958905249834,
1523
+ "num_tokens": 434995200.0,
1524
+ "step": 1660
1525
+ },
1526
+ {
1527
+ "epoch": 2.9772625947391886,
1528
+ "grad_norm": 0.5343450903892517,
1529
+ "learning_rate": 1.9178222137453656e-05,
1530
+ "loss": 1.141,
1531
+ "mean_token_accuracy": 0.7985142156481743,
1532
+ "num_tokens": 437616640.0,
1533
+ "step": 1670
1534
+ },
1535
+ {
1536
+ "epoch": 2.995095853767276,
1537
+ "grad_norm": 0.5583868026733398,
1538
+ "learning_rate": 1.8891489846552646e-05,
1539
+ "loss": 1.1432,
1540
+ "mean_token_accuracy": 0.7971971161663532,
1541
+ "num_tokens": 440238080.0,
1542
+ "step": 1680
1543
+ },
1544
+ {
1545
+ "epoch": 3.0,
1546
+ "eval_loss": 1.1234296560287476,
1547
+ "eval_mean_token_accuracy": 0.7894601821899414,
1548
+ "eval_num_tokens": 440942592.0,
1549
+ "eval_runtime": 2.9179,
1550
+ "eval_samples_per_second": 24.675,
1551
+ "eval_steps_per_second": 1.714,
1552
+ "step": 1683
1553
+ },
1554
+ {
1555
+ "epoch": 3.0124832813196614,
1556
+ "grad_norm": 0.5347409844398499,
1557
+ "learning_rate": 1.860560960887546e-05,
1558
+ "loss": 1.1253,
1559
+ "mean_token_accuracy": 0.8009015726737487,
1560
+ "num_tokens": 442777600.0,
1561
+ "step": 1690
1562
+ },
1563
+ {
1564
+ "epoch": 3.0303165403477483,
1565
+ "grad_norm": 0.5598998069763184,
1566
+ "learning_rate": 1.8320621300786688e-05,
1567
+ "loss": 1.106,
1568
+ "mean_token_accuracy": 0.8023651242256165,
1569
+ "num_tokens": 445399040.0,
1570
+ "step": 1700
1571
+ },
1572
+ {
1573
+ "epoch": 3.0481497993758357,
1574
+ "grad_norm": 0.5113139152526855,
1575
+ "learning_rate": 1.8036564674239043e-05,
1576
+ "loss": 1.0887,
1577
+ "mean_token_accuracy": 0.8040673099458218,
1578
+ "num_tokens": 448020480.0,
1579
+ "step": 1710
1580
+ },
1581
+ {
1582
+ "epoch": 3.065983058403923,
1583
+ "grad_norm": 0.44441869854927063,
1584
+ "learning_rate": 1.7753479351228444e-05,
1585
+ "loss": 1.1185,
1586
+ "mean_token_accuracy": 0.7999229028820991,
1587
+ "num_tokens": 450641920.0,
1588
+ "step": 1720
1589
+ },
1590
+ {
1591
+ "epoch": 3.0838163174320106,
1592
+ "grad_norm": 0.47339409589767456,
1593
+ "learning_rate": 1.7471404818267316e-05,
1594
+ "loss": 1.0925,
1595
+ "mean_token_accuracy": 0.8034070417284965,
1596
+ "num_tokens": 453263360.0,
1597
+ "step": 1730
1598
+ },
1599
+ {
1600
+ "epoch": 3.101649576460098,
1601
+ "grad_norm": 0.47856712341308594,
1602
+ "learning_rate": 1.7190380420876766e-05,
1603
+ "loss": 1.0911,
1604
+ "mean_token_accuracy": 0.8045188069343567,
1605
+ "num_tokens": 455884800.0,
1606
+ "step": 1740
1607
+ },
1608
+ {
1609
+ "epoch": 3.1194828354881854,
1610
+ "grad_norm": 0.4918275773525238,
1611
+ "learning_rate": 1.6910445358098415e-05,
1612
+ "loss": 1.1109,
1613
+ "mean_token_accuracy": 0.8015380732715129,
1614
+ "num_tokens": 458506240.0,
1615
+ "step": 1750
1616
+ },
1617
+ {
1618
+ "epoch": 3.137316094516273,
1619
+ "grad_norm": 0.5043804049491882,
1620
+ "learning_rate": 1.6631638677026674e-05,
1621
+ "loss": 1.1364,
1622
+ "mean_token_accuracy": 0.7974261090159416,
1623
+ "num_tokens": 461127680.0,
1624
+ "step": 1760
1625
+ },
1626
+ {
1627
+ "epoch": 3.15514935354436,
1628
+ "grad_norm": 0.4904772937297821,
1629
+ "learning_rate": 1.6353999267362237e-05,
1630
+ "loss": 1.0838,
1631
+ "mean_token_accuracy": 0.806386636942625,
1632
+ "num_tokens": 463749120.0,
1633
+ "step": 1770
1634
+ },
1635
+ {
1636
+ "epoch": 3.1729826125724476,
1637
+ "grad_norm": 0.48294153809547424,
1638
+ "learning_rate": 1.607756585598743e-05,
1639
+ "loss": 1.1389,
1640
+ "mean_token_accuracy": 0.7980314172804356,
1641
+ "num_tokens": 466370560.0,
1642
+ "step": 1780
1643
+ },
1644
+ {
1645
+ "epoch": 3.190815871600535,
1646
+ "grad_norm": 0.5139598250389099,
1647
+ "learning_rate": 1.580237700156443e-05,
1648
+ "loss": 1.1095,
1649
+ "mean_token_accuracy": 0.8025780878961086,
1650
+ "num_tokens": 468992000.0,
1651
+ "step": 1790
1652
+ },
1653
+ {
1654
+ "epoch": 3.2086491306286224,
1655
+ "grad_norm": 0.47298020124435425,
1656
+ "learning_rate": 1.5528471089156804e-05,
1657
+ "loss": 1.0931,
1658
+ "mean_token_accuracy": 0.8041585251688957,
1659
+ "num_tokens": 471613440.0,
1660
+ "step": 1800
1661
+ },
1662
+ {
1663
+ "epoch": 3.22648238965671,
1664
+ "grad_norm": 0.41903576254844666,
1665
+ "learning_rate": 1.5255886324875363e-05,
1666
+ "loss": 1.0954,
1667
+ "mean_token_accuracy": 0.8047371134161949,
1668
+ "num_tokens": 474234880.0,
1669
+ "step": 1810
1670
+ },
1671
+ {
1672
+ "epoch": 3.2443156486847973,
1673
+ "grad_norm": 0.4610045552253723,
1674
+ "learning_rate": 1.4984660730548856e-05,
1675
+ "loss": 1.1095,
1676
+ "mean_token_accuracy": 0.8022483348846435,
1677
+ "num_tokens": 476856320.0,
1678
+ "step": 1820
1679
+ },
1680
+ {
1681
+ "epoch": 3.2621489077128847,
1682
+ "grad_norm": 0.5105413198471069,
1683
+ "learning_rate": 1.4714832138420553e-05,
1684
+ "loss": 1.1258,
1685
+ "mean_token_accuracy": 0.7983997173607349,
1686
+ "num_tokens": 479477760.0,
1687
+ "step": 1830
1688
+ },
1689
+ {
1690
+ "epoch": 3.279982166740972,
1691
+ "grad_norm": 0.4393883943557739,
1692
+ "learning_rate": 1.4446438185871092e-05,
1693
+ "loss": 1.0566,
1694
+ "mean_token_accuracy": 0.8093036323785782,
1695
+ "num_tokens": 482099200.0,
1696
+ "step": 1840
1697
+ },
1698
+ {
1699
+ "epoch": 3.297815425769059,
1700
+ "grad_norm": 0.4618268311023712,
1701
+ "learning_rate": 1.4179516310168623e-05,
1702
+ "loss": 1.0697,
1703
+ "mean_token_accuracy": 0.8082189694046974,
1704
+ "num_tokens": 484720640.0,
1705
+ "step": 1850
1706
+ },
1707
+ {
1708
+ "epoch": 3.3156486847971465,
1709
+ "grad_norm": 0.47192856669425964,
1710
+ "learning_rate": 1.3914103743246757e-05,
1711
+ "loss": 1.1203,
1712
+ "mean_token_accuracy": 0.8002961635589599,
1713
+ "num_tokens": 487342080.0,
1714
+ "step": 1860
1715
+ },
1716
+ {
1717
+ "epoch": 3.333481943825234,
1718
+ "grad_norm": 0.4617849290370941,
1719
+ "learning_rate": 1.3650237506511331e-05,
1720
+ "loss": 1.0664,
1721
+ "mean_token_accuracy": 0.8089914366602897,
1722
+ "num_tokens": 489963520.0,
1723
+ "step": 1870
1724
+ },
1725
+ {
1726
+ "epoch": 3.3513152028533213,
1727
+ "grad_norm": 0.46263769268989563,
1728
+ "learning_rate": 1.3387954405676329e-05,
1729
+ "loss": 1.0855,
1730
+ "mean_token_accuracy": 0.8045512497425079,
1731
+ "num_tokens": 492584960.0,
1732
+ "step": 1880
1733
+ },
1734
+ {
1735
+ "epoch": 3.3691484618814087,
1736
+ "grad_norm": 0.47842440009117126,
1737
+ "learning_rate": 1.3127291025630073e-05,
1738
+ "loss": 1.0705,
1739
+ "mean_token_accuracy": 0.8085929870605468,
1740
+ "num_tokens": 495206400.0,
1741
+ "step": 1890
1742
+ },
1743
+ {
1744
+ "epoch": 3.386981720909496,
1745
+ "grad_norm": 0.43500861525535583,
1746
+ "learning_rate": 1.2868283725332028e-05,
1747
+ "loss": 1.0635,
1748
+ "mean_token_accuracy": 0.8091395162045956,
1749
+ "num_tokens": 497827840.0,
1750
+ "step": 1900
1751
+ },
1752
+ {
1753
+ "epoch": 3.4048149799375835,
1754
+ "grad_norm": 0.508797287940979,
1755
+ "learning_rate": 1.261096863274136e-05,
1756
+ "loss": 1.1052,
1757
+ "mean_token_accuracy": 0.8021300226449967,
1758
+ "num_tokens": 500449280.0,
1759
+ "step": 1910
1760
+ },
1761
+ {
1762
+ "epoch": 3.422648238965671,
1763
+ "grad_norm": 0.42378219962120056,
1764
+ "learning_rate": 1.2355381639777472e-05,
1765
+ "loss": 1.0936,
1766
+ "mean_token_accuracy": 0.8043497301638126,
1767
+ "num_tokens": 503070720.0,
1768
+ "step": 1920
1769
+ },
1770
+ {
1771
+ "epoch": 3.4404814979937584,
1772
+ "grad_norm": 0.4547331631183624,
1773
+ "learning_rate": 1.2101558397313634e-05,
1774
+ "loss": 1.1142,
1775
+ "mean_token_accuracy": 0.8015869230031967,
1776
+ "num_tokens": 505692160.0,
1777
+ "step": 1930
1778
+ },
1779
+ {
1780
+ "epoch": 3.4583147570218458,
1781
+ "grad_norm": 0.4319451153278351,
1782
+ "learning_rate": 1.1849534310204152e-05,
1783
+ "loss": 1.0835,
1784
+ "mean_token_accuracy": 0.8058400988578797,
1785
+ "num_tokens": 508313600.0,
1786
+ "step": 1940
1787
+ },
1788
+ {
1789
+ "epoch": 3.476148016049933,
1790
+ "grad_norm": 0.4666835069656372,
1791
+ "learning_rate": 1.1599344532345877e-05,
1792
+ "loss": 1.1072,
1793
+ "mean_token_accuracy": 0.801591893285513,
1794
+ "num_tokens": 510935040.0,
1795
+ "step": 1950
1796
+ },
1797
+ {
1798
+ "epoch": 3.4939812750780206,
1799
+ "grad_norm": 0.47252121567726135,
1800
+ "learning_rate": 1.1351023961774749e-05,
1801
+ "loss": 1.1245,
1802
+ "mean_token_accuracy": 0.7985180281102657,
1803
+ "num_tokens": 513556480.0,
1804
+ "step": 1960
1805
+ },
1806
+ {
1807
+ "epoch": 3.511814534106108,
1808
+ "grad_norm": 0.5152117013931274,
1809
+ "learning_rate": 1.110460723579796e-05,
1810
+ "loss": 1.0982,
1811
+ "mean_token_accuracy": 0.8033757455646992,
1812
+ "num_tokens": 516177920.0,
1813
+ "step": 1970
1814
+ },
1815
+ {
1816
+ "epoch": 3.5296477931341954,
1817
+ "grad_norm": 0.43357202410697937,
1818
+ "learning_rate": 1.0860128726162545e-05,
1819
+ "loss": 1.0632,
1820
+ "mean_token_accuracy": 0.8095081992447376,
1821
+ "num_tokens": 518799360.0,
1822
+ "step": 1980
1823
+ },
1824
+ {
1825
+ "epoch": 3.547481052162283,
1826
+ "grad_norm": 0.45368409156799316,
1827
+ "learning_rate": 1.0617622534260985e-05,
1828
+ "loss": 1.114,
1829
+ "mean_token_accuracy": 0.8014212884008884,
1830
+ "num_tokens": 521420800.0,
1831
+ "step": 1990
1832
+ },
1833
+ {
1834
+ "epoch": 3.5653143111903702,
1835
+ "grad_norm": 0.38939619064331055,
1836
+ "learning_rate": 1.0377122486374576e-05,
1837
+ "loss": 1.0449,
1838
+ "mean_token_accuracy": 0.8123023062944412,
1839
+ "num_tokens": 524042240.0,
1840
+ "step": 2000
1841
+ },
1842
+ {
1843
+ "epoch": 3.5831475702184576,
1844
+ "grad_norm": 0.42768579721450806,
1845
+ "learning_rate": 1.0138662128955053e-05,
1846
+ "loss": 1.104,
1847
+ "mean_token_accuracy": 0.8025334358215332,
1848
+ "num_tokens": 526663680.0,
1849
+ "step": 2010
1850
+ },
1851
+ {
1852
+ "epoch": 3.600980829246545,
1853
+ "grad_norm": 0.41337716579437256,
1854
+ "learning_rate": 9.902274723945381e-06,
1855
+ "loss": 1.0788,
1856
+ "mean_token_accuracy": 0.8076182380318642,
1857
+ "num_tokens": 529285120.0,
1858
+ "step": 2020
1859
+ },
1860
+ {
1861
+ "epoch": 3.618814088274632,
1862
+ "grad_norm": 0.4830559194087982,
1863
+ "learning_rate": 9.667993244140134e-06,
1864
+ "loss": 1.0967,
1865
+ "mean_token_accuracy": 0.8028597474098206,
1866
+ "num_tokens": 531906560.0,
1867
+ "step": 2030
1868
+ },
1869
+ {
1870
+ "epoch": 3.6366473473027194,
1871
+ "grad_norm": 0.4195765554904938,
1872
+ "learning_rate": 9.435850368586286e-06,
1873
+ "loss": 1.0773,
1874
+ "mean_token_accuracy": 0.8066488340497017,
1875
+ "num_tokens": 534528000.0,
1876
+ "step": 2040
1877
+ },
1878
+ {
1879
+ "epoch": 3.654480606330807,
1880
+ "grad_norm": 0.4251314103603363,
1881
+ "learning_rate": 9.20587847802488e-06,
1882
+ "loss": 1.0819,
1883
+ "mean_token_accuracy": 0.8052599862217903,
1884
+ "num_tokens": 537149440.0,
1885
+ "step": 2050
1886
+ },
1887
+ {
1888
+ "epoch": 3.6723138653588943,
1889
+ "grad_norm": 0.4114632308483124,
1890
+ "learning_rate": 8.978109650374397e-06,
1891
+ "loss": 1.0697,
1892
+ "mean_token_accuracy": 0.8077415138483047,
1893
+ "num_tokens": 539770880.0,
1894
+ "step": 2060
1895
+ },
1896
+ {
1897
+ "epoch": 3.6901471243869817,
1898
+ "grad_norm": 0.4304318428039551,
1899
+ "learning_rate": 8.752575656256321e-06,
1900
+ "loss": 1.1046,
1901
+ "mean_token_accuracy": 0.8016697481274605,
1902
+ "num_tokens": 542392320.0,
1903
+ "step": 2070
1904
+ },
1905
+ {
1906
+ "epoch": 3.707980383415069,
1907
+ "grad_norm": 0.47687673568725586,
1908
+ "learning_rate": 8.52930795456355e-06,
1909
+ "loss": 1.0558,
1910
+ "mean_token_accuracy": 0.8107806429266929,
1911
+ "num_tokens": 545013760.0,
1912
+ "step": 2080
1913
+ },
1914
+ {
1915
+ "epoch": 3.7258136424431565,
1916
+ "grad_norm": 0.45524051785469055,
1917
+ "learning_rate": 8.30833768807237e-06,
1918
+ "loss": 1.0816,
1919
+ "mean_token_accuracy": 0.8059305518865585,
1920
+ "num_tokens": 547635200.0,
1921
+ "step": 2090
1922
+ },
1923
+ {
1924
+ "epoch": 3.743646901471244,
1925
+ "grad_norm": 0.45478421449661255,
1926
+ "learning_rate": 8.089695679098375e-06,
1927
+ "loss": 1.0961,
1928
+ "mean_token_accuracy": 0.8029544025659561,
1929
+ "num_tokens": 550256640.0,
1930
+ "step": 2100
1931
+ },
1932
+ {
1933
+ "epoch": 3.7614801604993313,
1934
+ "grad_norm": 0.45910394191741943,
1935
+ "learning_rate": 7.87341242519728e-06,
1936
+ "loss": 1.0608,
1937
+ "mean_token_accuracy": 0.8082987323403359,
1938
+ "num_tokens": 552878080.0,
1939
+ "step": 2110
1940
+ },
1941
+ {
1942
+ "epoch": 3.7793134195274187,
1943
+ "grad_norm": 0.4360869824886322,
1944
+ "learning_rate": 7.659518094910839e-06,
1945
+ "loss": 1.0806,
1946
+ "mean_token_accuracy": 0.8052332706749439,
1947
+ "num_tokens": 555499520.0,
1948
+ "step": 2120
1949
+ },
1950
+ {
1951
+ "epoch": 3.797146678555506,
1952
+ "grad_norm": 0.46505963802337646,
1953
+ "learning_rate": 7.44804252355884e-06,
1954
+ "loss": 1.1143,
1955
+ "mean_token_accuracy": 0.8007854476571084,
1956
+ "num_tokens": 558120960.0,
1957
+ "step": 2130
1958
+ },
1959
+ {
1960
+ "epoch": 3.8149799375835935,
1961
+ "grad_norm": 0.40090060234069824,
1962
+ "learning_rate": 7.2390152090773765e-06,
1963
+ "loss": 1.079,
1964
+ "mean_token_accuracy": 0.8058068990707398,
1965
+ "num_tokens": 560742400.0,
1966
+ "step": 2140
1967
+ },
1968
+ {
1969
+ "epoch": 3.8328131966116805,
1970
+ "grad_norm": 0.41377872228622437,
1971
+ "learning_rate": 7.032465307904404e-06,
1972
+ "loss": 1.0496,
1973
+ "mean_token_accuracy": 0.8097444467246533,
1974
+ "num_tokens": 563363840.0,
1975
+ "step": 2150
1976
+ },
1977
+ {
1978
+ "epoch": 3.850646455639768,
1979
+ "grad_norm": 0.45236936211586,
1980
+ "learning_rate": 6.82842163091274e-06,
1981
+ "loss": 1.0819,
1982
+ "mean_token_accuracy": 0.8053412742912769,
1983
+ "num_tokens": 565985280.0,
1984
+ "step": 2160
1985
+ },
1986
+ {
1987
+ "epoch": 3.8684797146678553,
1988
+ "grad_norm": 0.42451992630958557,
1989
+ "learning_rate": 6.626912639391377e-06,
1990
+ "loss": 1.0649,
1991
+ "mean_token_accuracy": 0.8076308310031891,
1992
+ "num_tokens": 568606720.0,
1993
+ "step": 2170
1994
+ },
1995
+ {
1996
+ "epoch": 3.8863129736959428,
1997
+ "grad_norm": 0.43241724371910095,
1998
+ "learning_rate": 6.427966441075511e-06,
1999
+ "loss": 1.0821,
2000
+ "mean_token_accuracy": 0.805367611348629,
2001
+ "num_tokens": 571228160.0,
2002
+ "step": 2180
2003
+ },
2004
+ {
2005
+ "epoch": 3.90414623272403,
2006
+ "grad_norm": 0.41288912296295166,
2007
+ "learning_rate": 6.2316107862259026e-06,
2008
+ "loss": 1.1017,
2009
+ "mean_token_accuracy": 0.8022395603358745,
2010
+ "num_tokens": 573849600.0,
2011
+ "step": 2190
2012
+ },
2013
+ {
2014
+ "epoch": 3.9219794917521176,
2015
+ "grad_norm": 0.4336395263671875,
2016
+ "learning_rate": 6.037873063758112e-06,
2017
+ "loss": 1.0546,
2018
+ "mean_token_accuracy": 0.8097028434276581,
2019
+ "num_tokens": 576471040.0,
2020
+ "step": 2200
2021
+ },
2022
+ {
2023
+ "epoch": 3.939812750780205,
2024
+ "grad_norm": 0.4041893482208252,
2025
+ "learning_rate": 5.846780297422091e-06,
2026
+ "loss": 1.0603,
2027
+ "mean_token_accuracy": 0.8089009828865528,
2028
+ "num_tokens": 579092480.0,
2029
+ "step": 2210
2030
+ },
2031
+ {
2032
+ "epoch": 3.9576460098082924,
2033
+ "grad_norm": 0.39978188276290894,
2034
+ "learning_rate": 5.6583591420327684e-06,
2035
+ "loss": 1.0752,
2036
+ "mean_token_accuracy": 0.8069194242358207,
2037
+ "num_tokens": 581713920.0,
2038
+ "step": 2220
2039
+ },
2040
+ {
2041
+ "epoch": 3.97547926883638,
2042
+ "grad_norm": 0.41306570172309875,
2043
+ "learning_rate": 5.472635879752042e-06,
2044
+ "loss": 1.0757,
2045
+ "mean_token_accuracy": 0.8059477277100087,
2046
+ "num_tokens": 584335360.0,
2047
+ "step": 2230
2048
+ },
2049
+ {
2050
+ "epoch": 3.993312527864467,
2051
+ "grad_norm": 0.4413711428642273,
2052
+ "learning_rate": 5.289636416422822e-06,
2053
+ "loss": 1.1117,
2054
+ "mean_token_accuracy": 0.8009537614881992,
2055
+ "num_tokens": 586956800.0,
2056
+ "step": 2240
2057
+ },
2058
+ {
2059
+ "epoch": 4.0,
2060
+ "eval_loss": 1.0707859992980957,
2061
+ "eval_mean_token_accuracy": 0.7980337142944336,
2062
+ "eval_num_tokens": 587923456.0,
2063
+ "eval_runtime": 3.1887,
2064
+ "eval_samples_per_second": 22.579,
2065
+ "eval_steps_per_second": 1.568,
2066
+ "step": 2244
2067
+ },
2068
+ {
2069
+ "epoch": 4.010699955416852,
2070
+ "grad_norm": 0.3837050199508667,
2071
+ "learning_rate": 5.1093862779554776e-06,
2072
+ "loss": 1.0237,
2073
+ "mean_token_accuracy": 0.814745701276339,
2074
+ "num_tokens": 589496320.0,
2075
+ "step": 2250
2076
+ },
2077
+ {
2078
+ "epoch": 4.0285332144449395,
2079
+ "grad_norm": 0.39410483837127686,
2080
+ "learning_rate": 4.93191060676734e-06,
2081
+ "loss": 1.059,
2082
+ "mean_token_accuracy": 0.8092822574079037,
2083
+ "num_tokens": 592117760.0,
2084
+ "step": 2260
2085
+ },
2086
+ {
2087
+ "epoch": 4.046366473473027,
2088
+ "grad_norm": 0.4275108873844147,
2089
+ "learning_rate": 4.757234158275681e-06,
2090
+ "loss": 1.0876,
2091
+ "mean_token_accuracy": 0.8050832770764828,
2092
+ "num_tokens": 594739200.0,
2093
+ "step": 2270
2094
+ },
2095
+ {
2096
+ "epoch": 4.064199732501114,
2097
+ "grad_norm": 0.39579182863235474,
2098
+ "learning_rate": 4.585381297444688e-06,
2099
+ "loss": 1.0313,
2100
+ "mean_token_accuracy": 0.8126221284270286,
2101
+ "num_tokens": 597360640.0,
2102
+ "step": 2280
2103
+ },
2104
+ {
2105
+ "epoch": 4.082032991529202,
2106
+ "grad_norm": 0.4174662232398987,
2107
+ "learning_rate": 4.416375995386857e-06,
2108
+ "loss": 1.0677,
2109
+ "mean_token_accuracy": 0.808067063242197,
2110
+ "num_tokens": 599982080.0,
2111
+ "step": 2290
2112
+ },
2113
+ {
2114
+ "epoch": 4.099866250557289,
2115
+ "grad_norm": 0.426666796207428,
2116
+ "learning_rate": 4.2502418260193686e-06,
2117
+ "loss": 1.0551,
2118
+ "mean_token_accuracy": 0.8088761754333973,
2119
+ "num_tokens": 602603520.0,
2120
+ "step": 2300
2121
+ },
2122
+ {
2123
+ "epoch": 4.117699509585377,
2124
+ "grad_norm": 0.3763045370578766,
2125
+ "learning_rate": 4.087001962775849e-06,
2126
+ "loss": 1.0624,
2127
+ "mean_token_accuracy": 0.8099226720631123,
2128
+ "num_tokens": 605224960.0,
2129
+ "step": 2310
2130
+ },
2131
+ {
2132
+ "epoch": 4.135532768613464,
2133
+ "grad_norm": 0.40583643317222595,
2134
+ "learning_rate": 3.926679175373985e-06,
2135
+ "loss": 1.0593,
2136
+ "mean_token_accuracy": 0.8088261790573597,
2137
+ "num_tokens": 607846400.0,
2138
+ "step": 2320
2139
+ },
2140
+ {
2141
+ "epoch": 4.153366027641551,
2142
+ "grad_norm": 0.3805159330368042,
2143
+ "learning_rate": 3.7692958266394963e-06,
2144
+ "loss": 1.0686,
2145
+ "mean_token_accuracy": 0.808470857143402,
2146
+ "num_tokens": 610467840.0,
2147
+ "step": 2330
2148
+ },
2149
+ {
2150
+ "epoch": 4.171199286669639,
2151
+ "grad_norm": 0.36603036522865295,
2152
+ "learning_rate": 3.614873869386762e-06,
2153
+ "loss": 1.0475,
2154
+ "mean_token_accuracy": 0.8111096285283566,
2155
+ "num_tokens": 613089280.0,
2156
+ "step": 2340
2157
+ },
2158
+ {
2159
+ "epoch": 4.189032545697726,
2160
+ "grad_norm": 0.4105148911476135,
2161
+ "learning_rate": 3.463434843356783e-06,
2162
+ "loss": 1.0474,
2163
+ "mean_token_accuracy": 0.8102818146348,
2164
+ "num_tokens": 615710720.0,
2165
+ "step": 2350
2166
+ },
2167
+ {
2168
+ "epoch": 4.206865804725814,
2169
+ "grad_norm": 0.41211214661598206,
2170
+ "learning_rate": 3.314999872212618e-06,
2171
+ "loss": 1.1055,
2172
+ "mean_token_accuracy": 0.8027391448616982,
2173
+ "num_tokens": 618332160.0,
2174
+ "step": 2360
2175
+ },
2176
+ {
2177
+ "epoch": 4.224699063753901,
2178
+ "grad_norm": 0.409272164106369,
2179
+ "learning_rate": 3.169589660592978e-06,
2180
+ "loss": 1.0479,
2181
+ "mean_token_accuracy": 0.8102260909974575,
2182
+ "num_tokens": 620953600.0,
2183
+ "step": 2370
2184
+ },
2185
+ {
2186
+ "epoch": 4.242532322781988,
2187
+ "grad_norm": 0.4126708209514618,
2188
+ "learning_rate": 3.0272244912241453e-06,
2189
+ "loss": 1.0142,
2190
+ "mean_token_accuracy": 0.8164093054831028,
2191
+ "num_tokens": 623575040.0,
2192
+ "step": 2380
2193
+ }
2194
+ ],
2195
+ "logging_steps": 10,
2196
+ "max_steps": 2800,
2197
+ "num_input_tokens_seen": 0,
2198
+ "num_train_epochs": 5,
2199
+ "save_steps": 140,
2200
+ "stateful_callbacks": {
2201
+ "TrainerControl": {
2202
+ "args": {
2203
+ "should_epoch_stop": false,
2204
+ "should_evaluate": false,
2205
+ "should_log": false,
2206
+ "should_save": true,
2207
+ "should_training_stop": false
2208
+ },
2209
+ "attributes": {}
2210
+ }
2211
+ },
2212
+ "total_flos": 3.1048379184772547e+18,
2213
+ "train_batch_size": 16,
2214
+ "trial_name": null,
2215
+ "trial_params": null
2216
+ }
checkpoint-2380/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2520/config.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_configs": [
3
+ {
4
+ "vocab_size": 49152,
5
+ "max_position_embeddings": 8192,
6
+ "hidden_size": 960,
7
+ "intermediate_size": 2560,
8
+ "num_hidden_layers": 32,
9
+ "num_attention_heads": 15,
10
+ "num_key_value_heads": 5,
11
+ "hidden_act": "silu",
12
+ "initializer_range": 0.02,
13
+ "rms_norm_eps": 1e-05,
14
+ "pretraining_tp": 1,
15
+ "use_cache": true,
16
+ "rope_theta": 100000,
17
+ "rope_scaling": null,
18
+ "attention_bias": false,
19
+ "attention_dropout": 0.0,
20
+ "mlp_bias": false,
21
+ "head_dim": 64,
22
+ "return_dict": true,
23
+ "output_hidden_states": false,
24
+ "output_attentions": false,
25
+ "torchscript": false,
26
+ "torch_dtype": "float32",
27
+ "use_bfloat16": false,
28
+ "tf_legacy_loss": false,
29
+ "pruned_heads": {},
30
+ "tie_word_embeddings": true,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "is_decoder": false,
34
+ "cross_attention_hidden_size": null,
35
+ "add_cross_attention": false,
36
+ "tie_encoder_decoder": false,
37
+ "max_length": 20,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "early_stopping": false,
41
+ "num_beams": 1,
42
+ "num_beam_groups": 1,
43
+ "diversity_penalty": 0.0,
44
+ "temperature": 1.0,
45
+ "top_k": 50,
46
+ "top_p": 1.0,
47
+ "typical_p": 1.0,
48
+ "repetition_penalty": 1.0,
49
+ "length_penalty": 1.0,
50
+ "no_repeat_ngram_size": 0,
51
+ "encoder_no_repeat_ngram_size": 0,
52
+ "bad_words_ids": null,
53
+ "num_return_sequences": 1,
54
+ "output_scores": false,
55
+ "return_dict_in_generate": false,
56
+ "forced_bos_token_id": null,
57
+ "forced_eos_token_id": null,
58
+ "remove_invalid_values": false,
59
+ "exponential_decay_length_penalty": null,
60
+ "suppress_tokens": null,
61
+ "begin_suppress_tokens": null,
62
+ "architectures": [
63
+ "LlamaForCausalLM"
64
+ ],
65
+ "finetuning_task": null,
66
+ "id2label": {
67
+ "0": "LABEL_0",
68
+ "1": "LABEL_1"
69
+ },
70
+ "label2id": {
71
+ "LABEL_0": 0,
72
+ "LABEL_1": 1
73
+ },
74
+ "tokenizer_class": null,
75
+ "prefix": null,
76
+ "bos_token_id": 1,
77
+ "pad_token_id": 2,
78
+ "eos_token_id": 2,
79
+ "sep_token_id": null,
80
+ "decoder_start_token_id": null,
81
+ "task_specific_params": null,
82
+ "problem_type": null,
83
+ "_name_or_path": "HuggingFaceTB/SmolLM2-360M-Instruct",
84
+ "_attn_implementation_autoset": true,
85
+ "transformers_version": "4.51.3",
86
+ "is_llama_config": true,
87
+ "model_type": "llama",
88
+ "rope_interleaved": false,
89
+ "transformers.js_config": {
90
+ "kv_cache_dtype": {
91
+ "q4f16": "float16",
92
+ "fp16": "float16"
93
+ }
94
+ }
95
+ },
96
+ {
97
+ "vocab_size": 49152,
98
+ "max_position_embeddings": 8192,
99
+ "hidden_size": 576,
100
+ "intermediate_size": 1536,
101
+ "num_hidden_layers": 30,
102
+ "num_attention_heads": 9,
103
+ "num_key_value_heads": 3,
104
+ "hidden_act": "silu",
105
+ "initializer_range": 0.041666666666666664,
106
+ "rms_norm_eps": 1e-05,
107
+ "pretraining_tp": 1,
108
+ "use_cache": true,
109
+ "rope_theta": 100000,
110
+ "rope_scaling": null,
111
+ "attention_bias": false,
112
+ "attention_dropout": 0.0,
113
+ "mlp_bias": false,
114
+ "head_dim": 64,
115
+ "return_dict": true,
116
+ "output_hidden_states": false,
117
+ "output_attentions": false,
118
+ "torchscript": false,
119
+ "torch_dtype": "float32",
120
+ "use_bfloat16": false,
121
+ "tf_legacy_loss": false,
122
+ "pruned_heads": {},
123
+ "tie_word_embeddings": true,
124
+ "chunk_size_feed_forward": 0,
125
+ "is_encoder_decoder": false,
126
+ "is_decoder": false,
127
+ "cross_attention_hidden_size": null,
128
+ "add_cross_attention": false,
129
+ "tie_encoder_decoder": false,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "do_sample": false,
133
+ "early_stopping": false,
134
+ "num_beams": 1,
135
+ "num_beam_groups": 1,
136
+ "diversity_penalty": 0.0,
137
+ "temperature": 1.0,
138
+ "top_k": 50,
139
+ "top_p": 1.0,
140
+ "typical_p": 1.0,
141
+ "repetition_penalty": 1.0,
142
+ "length_penalty": 1.0,
143
+ "no_repeat_ngram_size": 0,
144
+ "encoder_no_repeat_ngram_size": 0,
145
+ "bad_words_ids": null,
146
+ "num_return_sequences": 1,
147
+ "output_scores": false,
148
+ "return_dict_in_generate": false,
149
+ "forced_bos_token_id": null,
150
+ "forced_eos_token_id": null,
151
+ "remove_invalid_values": false,
152
+ "exponential_decay_length_penalty": null,
153
+ "suppress_tokens": null,
154
+ "begin_suppress_tokens": null,
155
+ "architectures": [
156
+ "LlamaForCausalLM"
157
+ ],
158
+ "finetuning_task": null,
159
+ "id2label": {
160
+ "0": "LABEL_0",
161
+ "1": "LABEL_1"
162
+ },
163
+ "label2id": {
164
+ "LABEL_0": 0,
165
+ "LABEL_1": 1
166
+ },
167
+ "tokenizer_class": null,
168
+ "prefix": null,
169
+ "bos_token_id": 1,
170
+ "pad_token_id": 2,
171
+ "eos_token_id": 2,
172
+ "sep_token_id": null,
173
+ "decoder_start_token_id": null,
174
+ "task_specific_params": null,
175
+ "problem_type": null,
176
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct",
177
+ "_attn_implementation_autoset": true,
178
+ "transformers_version": "4.51.3",
179
+ "is_llama_config": true,
180
+ "model_type": "llama",
181
+ "rope_interleaved": false,
182
+ "transformers.js_config": {
183
+ "kv_cache_dtype": {
184
+ "q4f16": "float16",
185
+ "fp16": "float16"
186
+ }
187
+ }
188
+ }
189
+ ],
190
+ "num_models": 2
191
+ }
checkpoint-2520/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2520/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
checkpoint-2520/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2520/tokenizer_config.json ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<repo_name>",
30
+ "lstrip": false,
31
+ "normalized": false,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "4": {
37
+ "content": "<reponame>",
38
+ "lstrip": false,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ },
44
+ "5": {
45
+ "content": "<file_sep>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false,
50
+ "special": true
51
+ },
52
+ "6": {
53
+ "content": "<filename>",
54
+ "lstrip": false,
55
+ "normalized": false,
56
+ "rstrip": false,
57
+ "single_word": false,
58
+ "special": true
59
+ },
60
+ "7": {
61
+ "content": "<gh_stars>",
62
+ "lstrip": false,
63
+ "normalized": false,
64
+ "rstrip": false,
65
+ "single_word": false,
66
+ "special": true
67
+ },
68
+ "8": {
69
+ "content": "<issue_start>",
70
+ "lstrip": false,
71
+ "normalized": false,
72
+ "rstrip": false,
73
+ "single_word": false,
74
+ "special": true
75
+ },
76
+ "9": {
77
+ "content": "<issue_comment>",
78
+ "lstrip": false,
79
+ "normalized": false,
80
+ "rstrip": false,
81
+ "single_word": false,
82
+ "special": true
83
+ },
84
+ "10": {
85
+ "content": "<issue_closed>",
86
+ "lstrip": false,
87
+ "normalized": false,
88
+ "rstrip": false,
89
+ "single_word": false,
90
+ "special": true
91
+ },
92
+ "11": {
93
+ "content": "<jupyter_start>",
94
+ "lstrip": false,
95
+ "normalized": false,
96
+ "rstrip": false,
97
+ "single_word": false,
98
+ "special": true
99
+ },
100
+ "12": {
101
+ "content": "<jupyter_text>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false,
106
+ "special": true
107
+ },
108
+ "13": {
109
+ "content": "<jupyter_code>",
110
+ "lstrip": false,
111
+ "normalized": false,
112
+ "rstrip": false,
113
+ "single_word": false,
114
+ "special": true
115
+ },
116
+ "14": {
117
+ "content": "<jupyter_output>",
118
+ "lstrip": false,
119
+ "normalized": false,
120
+ "rstrip": false,
121
+ "single_word": false,
122
+ "special": true
123
+ },
124
+ "15": {
125
+ "content": "<jupyter_script>",
126
+ "lstrip": false,
127
+ "normalized": false,
128
+ "rstrip": false,
129
+ "single_word": false,
130
+ "special": true
131
+ },
132
+ "16": {
133
+ "content": "<empty_output>",
134
+ "lstrip": false,
135
+ "normalized": false,
136
+ "rstrip": false,
137
+ "single_word": false,
138
+ "special": true
139
+ }
140
+ },
141
+ "additional_special_tokens": [
142
+ "<|im_start|>",
143
+ "<|im_end|>"
144
+ ],
145
+ "bos_token": "<|im_start|>",
146
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful AI assistant named SmolLM, trained by Hugging Face<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
147
+ "clean_up_tokenization_spaces": false,
148
+ "eos_token": "<|im_end|>",
149
+ "extra_special_tokens": {},
150
+ "model_max_length": 8192,
151
+ "pad_token": "<|im_end|>",
152
+ "tokenizer_class": "GPT2Tokenizer",
153
+ "unk_token": "<|endoftext|>",
154
+ "vocab_size": 49152
155
+ }
checkpoint-2520/trainer_state.json ADDED
@@ -0,0 +1,2342 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 4.492197949175212,
6
+ "eval_steps": 500,
7
+ "global_step": 2520,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.017833259028087384,
14
+ "grad_norm": 33.452999114990234,
15
+ "learning_rate": 3.2142857142857143e-06,
16
+ "loss": 8.9944,
17
+ "mean_token_accuracy": 0.32701094649964946,
18
+ "num_tokens": 2621440.0,
19
+ "step": 10
20
+ },
21
+ {
22
+ "epoch": 0.03566651805617477,
23
+ "grad_norm": 2.596579074859619,
24
+ "learning_rate": 6.785714285714285e-06,
25
+ "loss": 4.9649,
26
+ "mean_token_accuracy": 0.5848119974136352,
27
+ "num_tokens": 5242880.0,
28
+ "step": 20
29
+ },
30
+ {
31
+ "epoch": 0.05349977708426215,
32
+ "grad_norm": 1.8979898691177368,
33
+ "learning_rate": 1.0357142857142859e-05,
34
+ "loss": 4.1262,
35
+ "mean_token_accuracy": 0.5906830102205276,
36
+ "num_tokens": 7864320.0,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.07133303611234953,
41
+ "grad_norm": 1.1842058897018433,
42
+ "learning_rate": 1.392857142857143e-05,
43
+ "loss": 3.7327,
44
+ "mean_token_accuracy": 0.608327355235815,
45
+ "num_tokens": 10485760.0,
46
+ "step": 40
47
+ },
48
+ {
49
+ "epoch": 0.08916629514043692,
50
+ "grad_norm": 0.922074556350708,
51
+ "learning_rate": 1.75e-05,
52
+ "loss": 3.4492,
53
+ "mean_token_accuracy": 0.6281662203371525,
54
+ "num_tokens": 13107200.0,
55
+ "step": 50
56
+ },
57
+ {
58
+ "epoch": 0.1069995541685243,
59
+ "grad_norm": 0.7813541889190674,
60
+ "learning_rate": 2.107142857142857e-05,
61
+ "loss": 3.1416,
62
+ "mean_token_accuracy": 0.648241326212883,
63
+ "num_tokens": 15728640.0,
64
+ "step": 60
65
+ },
66
+ {
67
+ "epoch": 0.12483281319661169,
68
+ "grad_norm": 1.1468125581741333,
69
+ "learning_rate": 2.4642857142857145e-05,
70
+ "loss": 2.8778,
71
+ "mean_token_accuracy": 0.6583510167896748,
72
+ "num_tokens": 18350080.0,
73
+ "step": 70
74
+ },
75
+ {
76
+ "epoch": 0.14266607222469907,
77
+ "grad_norm": 1.0490431785583496,
78
+ "learning_rate": 2.8214285714285714e-05,
79
+ "loss": 2.7108,
80
+ "mean_token_accuracy": 0.663086223602295,
81
+ "num_tokens": 20971520.0,
82
+ "step": 80
83
+ },
84
+ {
85
+ "epoch": 0.16049933125278645,
86
+ "grad_norm": 0.7813755869865417,
87
+ "learning_rate": 3.1785714285714284e-05,
88
+ "loss": 2.6039,
89
+ "mean_token_accuracy": 0.6650082431733608,
90
+ "num_tokens": 23592960.0,
91
+ "step": 90
92
+ },
93
+ {
94
+ "epoch": 0.17833259028087384,
95
+ "grad_norm": 0.6500654220581055,
96
+ "learning_rate": 3.5357142857142864e-05,
97
+ "loss": 2.4513,
98
+ "mean_token_accuracy": 0.674950385093689,
99
+ "num_tokens": 26214400.0,
100
+ "step": 100
101
+ },
102
+ {
103
+ "epoch": 0.19616584930896122,
104
+ "grad_norm": 0.81620192527771,
105
+ "learning_rate": 3.892857142857143e-05,
106
+ "loss": 2.3602,
107
+ "mean_token_accuracy": 0.6777540303766727,
108
+ "num_tokens": 28835840.0,
109
+ "step": 110
110
+ },
111
+ {
112
+ "epoch": 0.2139991083370486,
113
+ "grad_norm": 0.6131238341331482,
114
+ "learning_rate": 4.25e-05,
115
+ "loss": 2.3545,
116
+ "mean_token_accuracy": 0.6758881151676178,
117
+ "num_tokens": 31457280.0,
118
+ "step": 120
119
+ },
120
+ {
121
+ "epoch": 0.231832367365136,
122
+ "grad_norm": 0.9951689839363098,
123
+ "learning_rate": 4.607142857142857e-05,
124
+ "loss": 2.2624,
125
+ "mean_token_accuracy": 0.6806462205946445,
126
+ "num_tokens": 34078720.0,
127
+ "step": 130
128
+ },
129
+ {
130
+ "epoch": 0.24966562639322337,
131
+ "grad_norm": 0.7796569466590881,
132
+ "learning_rate": 4.964285714285715e-05,
133
+ "loss": 2.1505,
134
+ "mean_token_accuracy": 0.6918394304811954,
135
+ "num_tokens": 36700160.0,
136
+ "step": 140
137
+ },
138
+ {
139
+ "epoch": 0.26749888542131073,
140
+ "grad_norm": 0.8614610433578491,
141
+ "learning_rate": 4.9998587698517684e-05,
142
+ "loss": 2.1591,
143
+ "mean_token_accuracy": 0.68697941750288,
144
+ "num_tokens": 39321600.0,
145
+ "step": 150
146
+ },
147
+ {
148
+ "epoch": 0.28533214444939814,
149
+ "grad_norm": 0.6178464889526367,
150
+ "learning_rate": 4.9993705873562665e-05,
151
+ "loss": 2.1048,
152
+ "mean_token_accuracy": 0.6911555036902428,
153
+ "num_tokens": 41943040.0,
154
+ "step": 160
155
+ },
156
+ {
157
+ "epoch": 0.3031654034774855,
158
+ "grad_norm": 0.7255717515945435,
159
+ "learning_rate": 4.998533777009954e-05,
160
+ "loss": 2.0847,
161
+ "mean_token_accuracy": 0.6920825429260731,
162
+ "num_tokens": 44564480.0,
163
+ "step": 170
164
+ },
165
+ {
166
+ "epoch": 0.3209986625055729,
167
+ "grad_norm": 0.7800336480140686,
168
+ "learning_rate": 4.9973484555363726e-05,
169
+ "loss": 2.049,
170
+ "mean_token_accuracy": 0.6982546880841255,
171
+ "num_tokens": 47185920.0,
172
+ "step": 180
173
+ },
174
+ {
175
+ "epoch": 0.33883192153366026,
176
+ "grad_norm": 0.886745274066925,
177
+ "learning_rate": 4.995814788271582e-05,
178
+ "loss": 2.0301,
179
+ "mean_token_accuracy": 0.697366189956665,
180
+ "num_tokens": 49807360.0,
181
+ "step": 190
182
+ },
183
+ {
184
+ "epoch": 0.3566651805617477,
185
+ "grad_norm": 0.6190801858901978,
186
+ "learning_rate": 4.993932989141088e-05,
187
+ "loss": 1.9408,
188
+ "mean_token_accuracy": 0.7095360569655895,
189
+ "num_tokens": 52428800.0,
190
+ "step": 200
191
+ },
192
+ {
193
+ "epoch": 0.37449843958983503,
194
+ "grad_norm": 0.6333149671554565,
195
+ "learning_rate": 4.991703320630011e-05,
196
+ "loss": 1.9076,
197
+ "mean_token_accuracy": 0.7127362459897995,
198
+ "num_tokens": 55050240.0,
199
+ "step": 210
200
+ },
201
+ {
202
+ "epoch": 0.39233169861792244,
203
+ "grad_norm": 0.9046627879142761,
204
+ "learning_rate": 4.989126093746468e-05,
205
+ "loss": 1.8913,
206
+ "mean_token_accuracy": 0.7148448958992958,
207
+ "num_tokens": 57671680.0,
208
+ "step": 220
209
+ },
210
+ {
211
+ "epoch": 0.4101649576460098,
212
+ "grad_norm": 0.7778757214546204,
213
+ "learning_rate": 4.986201667978193e-05,
214
+ "loss": 1.9016,
215
+ "mean_token_accuracy": 0.7118340134620667,
216
+ "num_tokens": 60293120.0,
217
+ "step": 230
218
+ },
219
+ {
220
+ "epoch": 0.4279982166740972,
221
+ "grad_norm": 0.6346383690834045,
222
+ "learning_rate": 4.9829304512423944e-05,
223
+ "loss": 1.898,
224
+ "mean_token_accuracy": 0.7123855009675026,
225
+ "num_tokens": 62914560.0,
226
+ "step": 240
227
+ },
228
+ {
229
+ "epoch": 0.44583147570218457,
230
+ "grad_norm": 0.7308981418609619,
231
+ "learning_rate": 4.979312899828853e-05,
232
+ "loss": 1.8435,
233
+ "mean_token_accuracy": 0.7184088006615639,
234
+ "num_tokens": 65536000.0,
235
+ "step": 250
236
+ },
237
+ {
238
+ "epoch": 0.463664734730272,
239
+ "grad_norm": 0.6174861788749695,
240
+ "learning_rate": 4.9753495183362796e-05,
241
+ "loss": 1.8242,
242
+ "mean_token_accuracy": 0.7238672457635402,
243
+ "num_tokens": 68157440.0,
244
+ "step": 260
245
+ },
246
+ {
247
+ "epoch": 0.48149799375835933,
248
+ "grad_norm": 0.6507575511932373,
249
+ "learning_rate": 4.971040859601927e-05,
250
+ "loss": 1.791,
251
+ "mean_token_accuracy": 0.7251060992479325,
252
+ "num_tokens": 70778880.0,
253
+ "step": 270
254
+ },
255
+ {
256
+ "epoch": 0.49933125278644674,
257
+ "grad_norm": 0.8092382550239563,
258
+ "learning_rate": 4.966387524624482e-05,
259
+ "loss": 1.7879,
260
+ "mean_token_accuracy": 0.7255102708935738,
261
+ "num_tokens": 73400320.0,
262
+ "step": 280
263
+ },
264
+ {
265
+ "epoch": 0.5171645118145342,
266
+ "grad_norm": 0.5636172294616699,
267
+ "learning_rate": 4.9613901624802245e-05,
268
+ "loss": 1.7351,
269
+ "mean_token_accuracy": 0.7330663032829762,
270
+ "num_tokens": 76021760.0,
271
+ "step": 290
272
+ },
273
+ {
274
+ "epoch": 0.5349977708426215,
275
+ "grad_norm": 0.655262291431427,
276
+ "learning_rate": 4.956049470232503e-05,
277
+ "loss": 1.7464,
278
+ "mean_token_accuracy": 0.7302443347871304,
279
+ "num_tokens": 78643200.0,
280
+ "step": 300
281
+ },
282
+ {
283
+ "epoch": 0.5528310298707089,
284
+ "grad_norm": 0.9413332343101501,
285
+ "learning_rate": 4.9503661928344944e-05,
286
+ "loss": 1.7277,
287
+ "mean_token_accuracy": 0.730958030372858,
288
+ "num_tokens": 81264640.0,
289
+ "step": 310
290
+ },
291
+ {
292
+ "epoch": 0.5706642888987963,
293
+ "grad_norm": 0.7439999580383301,
294
+ "learning_rate": 4.9443411230252956e-05,
295
+ "loss": 1.6947,
296
+ "mean_token_accuracy": 0.7346593342721462,
297
+ "num_tokens": 83886080.0,
298
+ "step": 320
299
+ },
300
+ {
301
+ "epoch": 0.5884975479268837,
302
+ "grad_norm": 0.7665299773216248,
303
+ "learning_rate": 4.93797510121935e-05,
304
+ "loss": 1.7136,
305
+ "mean_token_accuracy": 0.7324949607253075,
306
+ "num_tokens": 86507520.0,
307
+ "step": 330
308
+ },
309
+ {
310
+ "epoch": 0.606330806954971,
311
+ "grad_norm": 0.7366372346878052,
312
+ "learning_rate": 4.931269015389217e-05,
313
+ "loss": 1.676,
314
+ "mean_token_accuracy": 0.7371297933161258,
315
+ "num_tokens": 89128960.0,
316
+ "step": 340
317
+ },
318
+ {
319
+ "epoch": 0.6241640659830584,
320
+ "grad_norm": 0.6571912169456482,
321
+ "learning_rate": 4.9242238009417175e-05,
322
+ "loss": 1.6252,
323
+ "mean_token_accuracy": 0.7434912383556366,
324
+ "num_tokens": 91750400.0,
325
+ "step": 350
326
+ },
327
+ {
328
+ "epoch": 0.6419973250111458,
329
+ "grad_norm": 0.7496657967567444,
330
+ "learning_rate": 4.9168404405874505e-05,
331
+ "loss": 1.657,
332
+ "mean_token_accuracy": 0.737925547361374,
333
+ "num_tokens": 94371840.0,
334
+ "step": 360
335
+ },
336
+ {
337
+ "epoch": 0.6598305840392331,
338
+ "grad_norm": 0.5275835394859314,
339
+ "learning_rate": 4.9091199642037244e-05,
340
+ "loss": 1.6723,
341
+ "mean_token_accuracy": 0.7358783438801766,
342
+ "num_tokens": 96993280.0,
343
+ "step": 370
344
+ },
345
+ {
346
+ "epoch": 0.6776638430673205,
347
+ "grad_norm": 0.6143308281898499,
348
+ "learning_rate": 4.901063448690901e-05,
349
+ "loss": 1.6292,
350
+ "mean_token_accuracy": 0.7412570238113403,
351
+ "num_tokens": 99614720.0,
352
+ "step": 380
353
+ },
354
+ {
355
+ "epoch": 0.6954971020954079,
356
+ "grad_norm": 0.860007643699646,
357
+ "learning_rate": 4.892672017822183e-05,
358
+ "loss": 1.6281,
359
+ "mean_token_accuracy": 0.7422710858285427,
360
+ "num_tokens": 102236160.0,
361
+ "step": 390
362
+ },
363
+ {
364
+ "epoch": 0.7133303611234953,
365
+ "grad_norm": 0.5295814275741577,
366
+ "learning_rate": 4.8839468420868606e-05,
367
+ "loss": 1.6112,
368
+ "mean_token_accuracy": 0.7446457460522652,
369
+ "num_tokens": 104857600.0,
370
+ "step": 400
371
+ },
372
+ {
373
+ "epoch": 0.7311636201515826,
374
+ "grad_norm": 0.7366165518760681,
375
+ "learning_rate": 4.8748891385270495e-05,
376
+ "loss": 1.6197,
377
+ "mean_token_accuracy": 0.7417318046092987,
378
+ "num_tokens": 107479040.0,
379
+ "step": 410
380
+ },
381
+ {
382
+ "epoch": 0.7489968791796701,
383
+ "grad_norm": 0.7033573985099792,
384
+ "learning_rate": 4.865500170567926e-05,
385
+ "loss": 1.5763,
386
+ "mean_token_accuracy": 0.7474127545952797,
387
+ "num_tokens": 110100480.0,
388
+ "step": 420
389
+ },
390
+ {
391
+ "epoch": 0.7668301382077575,
392
+ "grad_norm": 0.6527330875396729,
393
+ "learning_rate": 4.855781247841498e-05,
394
+ "loss": 1.5714,
395
+ "mean_token_accuracy": 0.7477886863052845,
396
+ "num_tokens": 112721920.0,
397
+ "step": 430
398
+ },
399
+ {
400
+ "epoch": 0.7846633972358449,
401
+ "grad_norm": 0.7344551682472229,
402
+ "learning_rate": 4.845733726003928e-05,
403
+ "loss": 1.5682,
404
+ "mean_token_accuracy": 0.7475516766309738,
405
+ "num_tokens": 115343360.0,
406
+ "step": 440
407
+ },
408
+ {
409
+ "epoch": 0.8024966562639322,
410
+ "grad_norm": 0.6840972304344177,
411
+ "learning_rate": 4.835359006546443e-05,
412
+ "loss": 1.5955,
413
+ "mean_token_accuracy": 0.7427977688610554,
414
+ "num_tokens": 117964800.0,
415
+ "step": 450
416
+ },
417
+ {
418
+ "epoch": 0.8203299152920196,
419
+ "grad_norm": 0.635522723197937,
420
+ "learning_rate": 4.824658536599836e-05,
421
+ "loss": 1.5459,
422
+ "mean_token_accuracy": 0.7526196874678135,
423
+ "num_tokens": 120586240.0,
424
+ "step": 460
425
+ },
426
+ {
427
+ "epoch": 0.838163174320107,
428
+ "grad_norm": 0.6874827146530151,
429
+ "learning_rate": 4.8136338087326216e-05,
430
+ "loss": 1.6108,
431
+ "mean_token_accuracy": 0.7413325920701027,
432
+ "num_tokens": 123207680.0,
433
+ "step": 470
434
+ },
435
+ {
436
+ "epoch": 0.8559964333481944,
437
+ "grad_norm": 0.6571274995803833,
438
+ "learning_rate": 4.8022863607428356e-05,
439
+ "loss": 1.539,
440
+ "mean_token_accuracy": 0.7521994858980179,
441
+ "num_tokens": 125829120.0,
442
+ "step": 480
443
+ },
444
+ {
445
+ "epoch": 0.8738296923762817,
446
+ "grad_norm": 0.8841336369514465,
447
+ "learning_rate": 4.790617775443537e-05,
448
+ "loss": 1.5486,
449
+ "mean_token_accuracy": 0.7492702730000019,
450
+ "num_tokens": 128450560.0,
451
+ "step": 490
452
+ },
453
+ {
454
+ "epoch": 0.8916629514043691,
455
+ "grad_norm": 0.690990149974823,
456
+ "learning_rate": 4.7786296804420274e-05,
457
+ "loss": 1.5646,
458
+ "mean_token_accuracy": 0.7455468386411667,
459
+ "num_tokens": 131072000.0,
460
+ "step": 500
461
+ },
462
+ {
463
+ "epoch": 0.9094962104324565,
464
+ "grad_norm": 0.7215222716331482,
465
+ "learning_rate": 4.766323747912818e-05,
466
+ "loss": 1.5085,
467
+ "mean_token_accuracy": 0.754824897646904,
468
+ "num_tokens": 133693440.0,
469
+ "step": 510
470
+ },
471
+ {
472
+ "epoch": 0.927329469460544,
473
+ "grad_norm": 0.6024391651153564,
474
+ "learning_rate": 4.753701694364392e-05,
475
+ "loss": 1.4856,
476
+ "mean_token_accuracy": 0.7570274345576763,
477
+ "num_tokens": 136314880.0,
478
+ "step": 520
479
+ },
480
+ {
481
+ "epoch": 0.9451627284886313,
482
+ "grad_norm": 0.6131455302238464,
483
+ "learning_rate": 4.740765280399767e-05,
484
+ "loss": 1.5433,
485
+ "mean_token_accuracy": 0.7495988786220551,
486
+ "num_tokens": 138936320.0,
487
+ "step": 530
488
+ },
489
+ {
490
+ "epoch": 0.9629959875167187,
491
+ "grad_norm": 0.7875815629959106,
492
+ "learning_rate": 4.72751631047092e-05,
493
+ "loss": 1.4917,
494
+ "mean_token_accuracy": 0.7559954360127449,
495
+ "num_tokens": 141557760.0,
496
+ "step": 540
497
+ },
498
+ {
499
+ "epoch": 0.9808292465448061,
500
+ "grad_norm": 0.7112794518470764,
501
+ "learning_rate": 4.71395663262709e-05,
502
+ "loss": 1.4784,
503
+ "mean_token_accuracy": 0.7572461232542992,
504
+ "num_tokens": 144179200.0,
505
+ "step": 550
506
+ },
507
+ {
508
+ "epoch": 0.9986625055728935,
509
+ "grad_norm": 0.6807007789611816,
510
+ "learning_rate": 4.7000881382570014e-05,
511
+ "loss": 1.5159,
512
+ "mean_token_accuracy": 0.7523536741733551,
513
+ "num_tokens": 146800640.0,
514
+ "step": 560
515
+ },
516
+ {
517
+ "epoch": 1.0,
518
+ "eval_loss": 1.494895339012146,
519
+ "eval_mean_token_accuracy": 0.7405715703964233,
520
+ "eval_num_tokens": 146980864.0,
521
+ "eval_runtime": 2.6056,
522
+ "eval_samples_per_second": 27.633,
523
+ "eval_steps_per_second": 1.919,
524
+ "step": 561
525
+ },
526
+ {
527
+ "epoch": 1.0160499331252786,
528
+ "grad_norm": 0.5480894446372986,
529
+ "learning_rate": 4.685912761825038e-05,
530
+ "loss": 1.4428,
531
+ "mean_token_accuracy": 0.7618689170250525,
532
+ "num_tokens": 149340160.0,
533
+ "step": 570
534
+ },
535
+ {
536
+ "epoch": 1.033883192153366,
537
+ "grad_norm": 0.6192435622215271,
538
+ "learning_rate": 4.671432480601416e-05,
539
+ "loss": 1.4711,
540
+ "mean_token_accuracy": 0.7556988887488842,
541
+ "num_tokens": 151961600.0,
542
+ "step": 580
543
+ },
544
+ {
545
+ "epoch": 1.0517164511814534,
546
+ "grad_norm": 0.5988848209381104,
547
+ "learning_rate": 4.656649314386378e-05,
548
+ "loss": 1.433,
549
+ "mean_token_accuracy": 0.7611474089324475,
550
+ "num_tokens": 154583040.0,
551
+ "step": 590
552
+ },
553
+ {
554
+ "epoch": 1.0695497102095408,
555
+ "grad_norm": 0.7051531672477722,
556
+ "learning_rate": 4.641565325228464e-05,
557
+ "loss": 1.415,
558
+ "mean_token_accuracy": 0.7643376767635346,
559
+ "num_tokens": 157204480.0,
560
+ "step": 600
561
+ },
562
+ {
563
+ "epoch": 1.0873829692376282,
564
+ "grad_norm": 0.6170579791069031,
565
+ "learning_rate": 4.6261826171368774e-05,
566
+ "loss": 1.4674,
567
+ "mean_token_accuracy": 0.7571297205984593,
568
+ "num_tokens": 159825920.0,
569
+ "step": 610
570
+ },
571
+ {
572
+ "epoch": 1.1052162282657156,
573
+ "grad_norm": 0.5691844820976257,
574
+ "learning_rate": 4.6105033357880104e-05,
575
+ "loss": 1.407,
576
+ "mean_token_accuracy": 0.7646078862249851,
577
+ "num_tokens": 162447360.0,
578
+ "step": 620
579
+ },
580
+ {
581
+ "epoch": 1.123049487293803,
582
+ "grad_norm": 0.5824171900749207,
583
+ "learning_rate": 4.594529668226146e-05,
584
+ "loss": 1.4185,
585
+ "mean_token_accuracy": 0.7614389970898628,
586
+ "num_tokens": 165068800.0,
587
+ "step": 630
588
+ },
589
+ {
590
+ "epoch": 1.1408827463218902,
591
+ "grad_norm": 0.5765758156776428,
592
+ "learning_rate": 4.578263842558402e-05,
593
+ "loss": 1.42,
594
+ "mean_token_accuracy": 0.7633033894002438,
595
+ "num_tokens": 167690240.0,
596
+ "step": 640
597
+ },
598
+ {
599
+ "epoch": 1.1587160053499777,
600
+ "grad_norm": 0.5261256694793701,
601
+ "learning_rate": 4.5617081276439355e-05,
602
+ "loss": 1.4384,
603
+ "mean_token_accuracy": 0.7604428693652153,
604
+ "num_tokens": 170311680.0,
605
+ "step": 650
606
+ },
607
+ {
608
+ "epoch": 1.176549264378065,
609
+ "grad_norm": 0.913208544254303,
610
+ "learning_rate": 4.544864832777469e-05,
611
+ "loss": 1.4775,
612
+ "mean_token_accuracy": 0.754877183586359,
613
+ "num_tokens": 172933120.0,
614
+ "step": 660
615
+ },
616
+ {
617
+ "epoch": 1.1943825234061525,
618
+ "grad_norm": 0.5799976587295532,
619
+ "learning_rate": 4.527736307367179e-05,
620
+ "loss": 1.4387,
621
+ "mean_token_accuracy": 0.7589173935353756,
622
+ "num_tokens": 175554560.0,
623
+ "step": 670
624
+ },
625
+ {
626
+ "epoch": 1.2122157824342399,
627
+ "grad_norm": 0.635047435760498,
628
+ "learning_rate": 4.510324940606979e-05,
629
+ "loss": 1.4439,
630
+ "mean_token_accuracy": 0.7597284145653248,
631
+ "num_tokens": 178176000.0,
632
+ "step": 680
633
+ },
634
+ {
635
+ "epoch": 1.2300490414623273,
636
+ "grad_norm": 0.5884873270988464,
637
+ "learning_rate": 4.4926331611432687e-05,
638
+ "loss": 1.3835,
639
+ "mean_token_accuracy": 0.7659787967801094,
640
+ "num_tokens": 180797440.0,
641
+ "step": 690
642
+ },
643
+ {
644
+ "epoch": 1.2478823004904147,
645
+ "grad_norm": 0.7232898473739624,
646
+ "learning_rate": 4.4746634367361634e-05,
647
+ "loss": 1.3629,
648
+ "mean_token_accuracy": 0.7691854700446129,
649
+ "num_tokens": 183418880.0,
650
+ "step": 700
651
+ },
652
+ {
653
+ "epoch": 1.265715559518502,
654
+ "grad_norm": 0.6569238305091858,
655
+ "learning_rate": 4.4564182739152805e-05,
656
+ "loss": 1.397,
657
+ "mean_token_accuracy": 0.7651830442249775,
658
+ "num_tokens": 186040320.0,
659
+ "step": 710
660
+ },
661
+ {
662
+ "epoch": 1.2835488185465893,
663
+ "grad_norm": 0.8630965352058411,
664
+ "learning_rate": 4.437900217630109e-05,
665
+ "loss": 1.4256,
666
+ "mean_token_accuracy": 0.7605882868170738,
667
+ "num_tokens": 188661760.0,
668
+ "step": 720
669
+ },
670
+ {
671
+ "epoch": 1.3013820775746767,
672
+ "grad_norm": 0.7600955367088318,
673
+ "learning_rate": 4.419111850895028e-05,
674
+ "loss": 1.3975,
675
+ "mean_token_accuracy": 0.7634797111153603,
676
+ "num_tokens": 191283200.0,
677
+ "step": 730
678
+ },
679
+ {
680
+ "epoch": 1.3192153366027641,
681
+ "grad_norm": 0.6687177419662476,
682
+ "learning_rate": 4.4000557944290086e-05,
683
+ "loss": 1.3616,
684
+ "mean_token_accuracy": 0.7703544825315476,
685
+ "num_tokens": 193904640.0,
686
+ "step": 740
687
+ },
688
+ {
689
+ "epoch": 1.3370485956308515,
690
+ "grad_norm": 0.719222903251648,
691
+ "learning_rate": 4.3807347062900624e-05,
692
+ "loss": 1.3788,
693
+ "mean_token_accuracy": 0.7648544363677502,
694
+ "num_tokens": 196526080.0,
695
+ "step": 750
696
+ },
697
+ {
698
+ "epoch": 1.354881854658939,
699
+ "grad_norm": 0.5986538529396057,
700
+ "learning_rate": 4.361151281504474e-05,
701
+ "loss": 1.364,
702
+ "mean_token_accuracy": 0.770479665696621,
703
+ "num_tokens": 199147520.0,
704
+ "step": 760
705
+ },
706
+ {
707
+ "epoch": 1.3727151136870264,
708
+ "grad_norm": 0.6472992897033691,
709
+ "learning_rate": 4.3413082516908893e-05,
710
+ "loss": 1.4045,
711
+ "mean_token_accuracy": 0.7644098080694676,
712
+ "num_tokens": 201768960.0,
713
+ "step": 770
714
+ },
715
+ {
716
+ "epoch": 1.3905483727151138,
717
+ "grad_norm": 0.44522207975387573,
718
+ "learning_rate": 4.321208384679285e-05,
719
+ "loss": 1.3314,
720
+ "mean_token_accuracy": 0.773588253557682,
721
+ "num_tokens": 204390400.0,
722
+ "step": 780
723
+ },
724
+ {
725
+ "epoch": 1.4083816317432012,
726
+ "grad_norm": 0.5019171833992004,
727
+ "learning_rate": 4.3008544841248986e-05,
728
+ "loss": 1.3591,
729
+ "mean_token_accuracy": 0.7695087313652038,
730
+ "num_tokens": 207011840.0,
731
+ "step": 790
732
+ },
733
+ {
734
+ "epoch": 1.4262148907712884,
735
+ "grad_norm": 0.7287567257881165,
736
+ "learning_rate": 4.280249389117157e-05,
737
+ "loss": 1.3773,
738
+ "mean_token_accuracy": 0.7675351880490779,
739
+ "num_tokens": 209633280.0,
740
+ "step": 800
741
+ },
742
+ {
743
+ "epoch": 1.4440481497993758,
744
+ "grad_norm": 1.0294618606567383,
745
+ "learning_rate": 4.2593959737836604e-05,
746
+ "loss": 1.3719,
747
+ "mean_token_accuracy": 0.7668005004525185,
748
+ "num_tokens": 212254720.0,
749
+ "step": 810
750
+ },
751
+ {
752
+ "epoch": 1.4618814088274632,
753
+ "grad_norm": 0.6181490421295166,
754
+ "learning_rate": 4.2382971468892806e-05,
755
+ "loss": 1.3168,
756
+ "mean_token_accuracy": 0.7758259050548076,
757
+ "num_tokens": 214876160.0,
758
+ "step": 820
759
+ },
760
+ {
761
+ "epoch": 1.4797146678555506,
762
+ "grad_norm": 0.5945810079574585,
763
+ "learning_rate": 4.216955851430432e-05,
764
+ "loss": 1.3536,
765
+ "mean_token_accuracy": 0.7710937492549419,
766
+ "num_tokens": 217497600.0,
767
+ "step": 830
768
+ },
769
+ {
770
+ "epoch": 1.497547926883638,
771
+ "grad_norm": 0.6803934574127197,
772
+ "learning_rate": 4.195375064224562e-05,
773
+ "loss": 1.324,
774
+ "mean_token_accuracy": 0.7753530308604241,
775
+ "num_tokens": 220119040.0,
776
+ "step": 840
777
+ },
778
+ {
779
+ "epoch": 1.5153811859117252,
780
+ "grad_norm": 0.5748502612113953,
781
+ "learning_rate": 4.1735577954949225e-05,
782
+ "loss": 1.3057,
783
+ "mean_token_accuracy": 0.7755480580031872,
784
+ "num_tokens": 222740480.0,
785
+ "step": 850
786
+ },
787
+ {
788
+ "epoch": 1.5332144449398126,
789
+ "grad_norm": 0.5675086379051208,
790
+ "learning_rate": 4.151507088450692e-05,
791
+ "loss": 1.302,
792
+ "mean_token_accuracy": 0.7765964694321156,
793
+ "num_tokens": 225361920.0,
794
+ "step": 860
795
+ },
796
+ {
797
+ "epoch": 1.5510477039679,
798
+ "grad_norm": 0.6266924142837524,
799
+ "learning_rate": 4.129226018862484e-05,
800
+ "loss": 1.353,
801
+ "mean_token_accuracy": 0.770804837346077,
802
+ "num_tokens": 227983360.0,
803
+ "step": 870
804
+ },
805
+ {
806
+ "epoch": 1.5688809629959874,
807
+ "grad_norm": 0.551527202129364,
808
+ "learning_rate": 4.106717694633321e-05,
809
+ "loss": 1.3477,
810
+ "mean_token_accuracy": 0.77206010222435,
811
+ "num_tokens": 230604800.0,
812
+ "step": 880
813
+ },
814
+ {
815
+ "epoch": 1.5867142220240749,
816
+ "grad_norm": 0.6628434658050537,
817
+ "learning_rate": 4.0839852553651265e-05,
818
+ "loss": 1.3244,
819
+ "mean_token_accuracy": 0.773486353456974,
820
+ "num_tokens": 233226240.0,
821
+ "step": 890
822
+ },
823
+ {
824
+ "epoch": 1.6045474810521623,
825
+ "grad_norm": 0.5956223011016846,
826
+ "learning_rate": 4.0610318719207906e-05,
827
+ "loss": 1.3298,
828
+ "mean_token_accuracy": 0.7733359761536122,
829
+ "num_tokens": 235847680.0,
830
+ "step": 900
831
+ },
832
+ {
833
+ "epoch": 1.6223807400802497,
834
+ "grad_norm": 0.6573621034622192,
835
+ "learning_rate": 4.037860745981881e-05,
836
+ "loss": 1.283,
837
+ "mean_token_accuracy": 0.7787142746150494,
838
+ "num_tokens": 238469120.0,
839
+ "step": 910
840
+ },
841
+ {
842
+ "epoch": 1.640213999108337,
843
+ "grad_norm": 0.607742965221405,
844
+ "learning_rate": 4.01447510960205e-05,
845
+ "loss": 1.2955,
846
+ "mean_token_accuracy": 0.7780937045812607,
847
+ "num_tokens": 241090560.0,
848
+ "step": 920
849
+ },
850
+ {
851
+ "epoch": 1.6580472581364245,
852
+ "grad_norm": 0.6278334856033325,
853
+ "learning_rate": 3.99087822475621e-05,
854
+ "loss": 1.3215,
855
+ "mean_token_accuracy": 0.7741382211446762,
856
+ "num_tokens": 243712000.0,
857
+ "step": 930
858
+ },
859
+ {
860
+ "epoch": 1.675880517164512,
861
+ "grad_norm": 0.5402811169624329,
862
+ "learning_rate": 3.96707338288553e-05,
863
+ "loss": 1.3113,
864
+ "mean_token_accuracy": 0.7751454092562199,
865
+ "num_tokens": 246333440.0,
866
+ "step": 940
867
+ },
868
+ {
869
+ "epoch": 1.6937137761925993,
870
+ "grad_norm": 0.6165571212768555,
871
+ "learning_rate": 3.9430639044383274e-05,
872
+ "loss": 1.3093,
873
+ "mean_token_accuracy": 0.7768628627061844,
874
+ "num_tokens": 248954880.0,
875
+ "step": 950
876
+ },
877
+ {
878
+ "epoch": 1.7115470352206867,
879
+ "grad_norm": 0.6086046695709229,
880
+ "learning_rate": 3.9188531384069096e-05,
881
+ "loss": 1.3125,
882
+ "mean_token_accuracy": 0.7762621335685254,
883
+ "num_tokens": 251576320.0,
884
+ "step": 960
885
+ },
886
+ {
887
+ "epoch": 1.729380294248774,
888
+ "grad_norm": 0.682686984539032,
889
+ "learning_rate": 3.8944444618604356e-05,
890
+ "loss": 1.2821,
891
+ "mean_token_accuracy": 0.7795726217329502,
892
+ "num_tokens": 254197760.0,
893
+ "step": 970
894
+ },
895
+ {
896
+ "epoch": 1.7472135532768613,
897
+ "grad_norm": 0.5615099668502808,
898
+ "learning_rate": 3.869841279473862e-05,
899
+ "loss": 1.3095,
900
+ "mean_token_accuracy": 0.7766918800771236,
901
+ "num_tokens": 256819200.0,
902
+ "step": 980
903
+ },
904
+ {
905
+ "epoch": 1.7650468123049488,
906
+ "grad_norm": 0.5341463088989258,
907
+ "learning_rate": 3.845047023053037e-05,
908
+ "loss": 1.2864,
909
+ "mean_token_accuracy": 0.779373399913311,
910
+ "num_tokens": 259440640.0,
911
+ "step": 990
912
+ },
913
+ {
914
+ "epoch": 1.7828800713330362,
915
+ "grad_norm": 0.5641157627105713,
916
+ "learning_rate": 3.82006515105601e-05,
917
+ "loss": 1.2857,
918
+ "mean_token_accuracy": 0.7798695519566536,
919
+ "num_tokens": 262062080.0,
920
+ "step": 1000
921
+ },
922
+ {
923
+ "epoch": 1.8007133303611234,
924
+ "grad_norm": 0.5196830630302429,
925
+ "learning_rate": 3.794899148110628e-05,
926
+ "loss": 1.2949,
927
+ "mean_token_accuracy": 0.7788909852504731,
928
+ "num_tokens": 264683520.0,
929
+ "step": 1010
930
+ },
931
+ {
932
+ "epoch": 1.8185465893892108,
933
+ "grad_norm": 0.4990008771419525,
934
+ "learning_rate": 3.7695525245284716e-05,
935
+ "loss": 1.254,
936
+ "mean_token_accuracy": 0.7839147254824639,
937
+ "num_tokens": 267304960.0,
938
+ "step": 1020
939
+ },
940
+ {
941
+ "epoch": 1.8363798484172982,
942
+ "grad_norm": 0.5643882751464844,
943
+ "learning_rate": 3.7440288158152187e-05,
944
+ "loss": 1.2446,
945
+ "mean_token_accuracy": 0.785726061463356,
946
+ "num_tokens": 269926400.0,
947
+ "step": 1030
948
+ },
949
+ {
950
+ "epoch": 1.8542131074453856,
951
+ "grad_norm": 0.8107313513755798,
952
+ "learning_rate": 3.7183315821774914e-05,
953
+ "loss": 1.2974,
954
+ "mean_token_accuracy": 0.7775464087724686,
955
+ "num_tokens": 272547840.0,
956
+ "step": 1040
957
+ },
958
+ {
959
+ "epoch": 1.872046366473473,
960
+ "grad_norm": 0.5176706314086914,
961
+ "learning_rate": 3.692464408026253e-05,
962
+ "loss": 1.2916,
963
+ "mean_token_accuracy": 0.7772151328623295,
964
+ "num_tokens": 275169280.0,
965
+ "step": 1050
966
+ },
967
+ {
968
+ "epoch": 1.8898796255015604,
969
+ "grad_norm": 0.5898165106773376,
970
+ "learning_rate": 3.666430901476833e-05,
971
+ "loss": 1.2716,
972
+ "mean_token_accuracy": 0.7806546173989772,
973
+ "num_tokens": 277790720.0,
974
+ "step": 1060
975
+ },
976
+ {
977
+ "epoch": 1.9077128845296478,
978
+ "grad_norm": 0.5368467569351196,
979
+ "learning_rate": 3.640234693845645e-05,
980
+ "loss": 1.2491,
981
+ "mean_token_accuracy": 0.7844746112823486,
982
+ "num_tokens": 280412160.0,
983
+ "step": 1070
984
+ },
985
+ {
986
+ "epoch": 1.9255461435577352,
987
+ "grad_norm": 0.5310569405555725,
988
+ "learning_rate": 3.613879439143667e-05,
989
+ "loss": 1.2825,
990
+ "mean_token_accuracy": 0.7795199528336525,
991
+ "num_tokens": 283033600.0,
992
+ "step": 1080
993
+ },
994
+ {
995
+ "epoch": 1.9433794025858226,
996
+ "grad_norm": 0.6114910840988159,
997
+ "learning_rate": 3.58736881356676e-05,
998
+ "loss": 1.2595,
999
+ "mean_token_accuracy": 0.7831220239400863,
1000
+ "num_tokens": 285655040.0,
1001
+ "step": 1090
1002
+ },
1003
+ {
1004
+ "epoch": 1.96121266161391,
1005
+ "grad_norm": 0.5786811113357544,
1006
+ "learning_rate": 3.5607065149828843e-05,
1007
+ "loss": 1.2694,
1008
+ "mean_token_accuracy": 0.7822808526456356,
1009
+ "num_tokens": 288276480.0,
1010
+ "step": 1100
1011
+ },
1012
+ {
1013
+ "epoch": 1.9790459206419975,
1014
+ "grad_norm": 0.6175279021263123,
1015
+ "learning_rate": 3.533896262416302e-05,
1016
+ "loss": 1.2712,
1017
+ "mean_token_accuracy": 0.7803981445729733,
1018
+ "num_tokens": 290897920.0,
1019
+ "step": 1110
1020
+ },
1021
+ {
1022
+ "epoch": 1.9968791796700847,
1023
+ "grad_norm": 0.568803608417511,
1024
+ "learning_rate": 3.506941795528821e-05,
1025
+ "loss": 1.2576,
1026
+ "mean_token_accuracy": 0.7821117818355561,
1027
+ "num_tokens": 293519360.0,
1028
+ "step": 1120
1029
+ },
1030
+ {
1031
+ "epoch": 2.0,
1032
+ "eval_loss": 1.2395988702774048,
1033
+ "eval_mean_token_accuracy": 0.7735649704933166,
1034
+ "eval_num_tokens": 293961728.0,
1035
+ "eval_runtime": 2.8804,
1036
+ "eval_samples_per_second": 24.997,
1037
+ "eval_steps_per_second": 1.736,
1038
+ "step": 1122
1039
+ },
1040
+ {
1041
+ "epoch": 2.0142666072224698,
1042
+ "grad_norm": 0.6652095913887024,
1043
+ "learning_rate": 3.479846874098167e-05,
1044
+ "loss": 1.2741,
1045
+ "mean_token_accuracy": 0.7794669828353784,
1046
+ "num_tokens": 296058880.0,
1047
+ "step": 1130
1048
+ },
1049
+ {
1050
+ "epoch": 2.032099866250557,
1051
+ "grad_norm": 0.7581707835197449,
1052
+ "learning_rate": 3.452615277493544e-05,
1053
+ "loss": 1.199,
1054
+ "mean_token_accuracy": 0.7900658771395683,
1055
+ "num_tokens": 298680320.0,
1056
+ "step": 1140
1057
+ },
1058
+ {
1059
+ "epoch": 2.0499331252786446,
1060
+ "grad_norm": 0.5383137464523315,
1061
+ "learning_rate": 3.425250804148464e-05,
1062
+ "loss": 1.2304,
1063
+ "mean_token_accuracy": 0.7861943542957306,
1064
+ "num_tokens": 301301760.0,
1065
+ "step": 1150
1066
+ },
1067
+ {
1068
+ "epoch": 2.067766384306732,
1069
+ "grad_norm": 0.5952672958374023,
1070
+ "learning_rate": 3.3977572710309166e-05,
1071
+ "loss": 1.2445,
1072
+ "mean_token_accuracy": 0.7836834371089936,
1073
+ "num_tokens": 303923200.0,
1074
+ "step": 1160
1075
+ },
1076
+ {
1077
+ "epoch": 2.0855996433348194,
1078
+ "grad_norm": 0.646511435508728,
1079
+ "learning_rate": 3.3701385131109616e-05,
1080
+ "loss": 1.2019,
1081
+ "mean_token_accuracy": 0.7892128698527813,
1082
+ "num_tokens": 306544640.0,
1083
+ "step": 1170
1084
+ },
1085
+ {
1086
+ "epoch": 2.103432902362907,
1087
+ "grad_norm": 0.514665961265564,
1088
+ "learning_rate": 3.342398382825791e-05,
1089
+ "loss": 1.2614,
1090
+ "mean_token_accuracy": 0.7812080189585686,
1091
+ "num_tokens": 309166080.0,
1092
+ "step": 1180
1093
+ },
1094
+ {
1095
+ "epoch": 2.121266161390994,
1096
+ "grad_norm": 0.5490918159484863,
1097
+ "learning_rate": 3.31454074954238e-05,
1098
+ "loss": 1.2108,
1099
+ "mean_token_accuracy": 0.7876022845506668,
1100
+ "num_tokens": 311787520.0,
1101
+ "step": 1190
1102
+ },
1103
+ {
1104
+ "epoch": 2.1390994204190816,
1105
+ "grad_norm": 0.6549789309501648,
1106
+ "learning_rate": 3.286569499017752e-05,
1107
+ "loss": 1.2024,
1108
+ "mean_token_accuracy": 0.7898712314665317,
1109
+ "num_tokens": 314408960.0,
1110
+ "step": 1200
1111
+ },
1112
+ {
1113
+ "epoch": 2.156932679447169,
1114
+ "grad_norm": 0.5743107199668884,
1115
+ "learning_rate": 3.2584885328569785e-05,
1116
+ "loss": 1.1835,
1117
+ "mean_token_accuracy": 0.7928275369107723,
1118
+ "num_tokens": 317030400.0,
1119
+ "step": 1210
1120
+ },
1121
+ {
1122
+ "epoch": 2.1747659384752565,
1123
+ "grad_norm": 0.6201761960983276,
1124
+ "learning_rate": 3.2303017679689525e-05,
1125
+ "loss": 1.1939,
1126
+ "mean_token_accuracy": 0.7908047571778297,
1127
+ "num_tokens": 319651840.0,
1128
+ "step": 1220
1129
+ },
1130
+ {
1131
+ "epoch": 2.192599197503344,
1132
+ "grad_norm": 0.550830602645874,
1133
+ "learning_rate": 3.2020131360200354e-05,
1134
+ "loss": 1.1864,
1135
+ "mean_token_accuracy": 0.7930233284831048,
1136
+ "num_tokens": 322273280.0,
1137
+ "step": 1230
1138
+ },
1139
+ {
1140
+ "epoch": 2.2104324565314313,
1141
+ "grad_norm": 0.5972590446472168,
1142
+ "learning_rate": 3.173626582885645e-05,
1143
+ "loss": 1.2177,
1144
+ "mean_token_accuracy": 0.7885610058903694,
1145
+ "num_tokens": 324894720.0,
1146
+ "step": 1240
1147
+ },
1148
+ {
1149
+ "epoch": 2.2282657155595187,
1150
+ "grad_norm": 0.5069220066070557,
1151
+ "learning_rate": 3.145146068099858e-05,
1152
+ "loss": 1.2279,
1153
+ "mean_token_accuracy": 0.7862870991230011,
1154
+ "num_tokens": 327516160.0,
1155
+ "step": 1250
1156
+ },
1157
+ {
1158
+ "epoch": 2.246098974587606,
1159
+ "grad_norm": 0.5100940465927124,
1160
+ "learning_rate": 3.116575564303109e-05,
1161
+ "loss": 1.1777,
1162
+ "mean_token_accuracy": 0.7929233327507973,
1163
+ "num_tokens": 330137600.0,
1164
+ "step": 1260
1165
+ },
1166
+ {
1167
+ "epoch": 2.2639322336156935,
1168
+ "grad_norm": 0.6175646781921387,
1169
+ "learning_rate": 3.087919056688067e-05,
1170
+ "loss": 1.173,
1171
+ "mean_token_accuracy": 0.7950430497527122,
1172
+ "num_tokens": 332759040.0,
1173
+ "step": 1270
1174
+ },
1175
+ {
1176
+ "epoch": 2.2817654926437805,
1177
+ "grad_norm": 0.5933928489685059,
1178
+ "learning_rate": 3.059180542443746e-05,
1179
+ "loss": 1.2057,
1180
+ "mean_token_accuracy": 0.7882316343486309,
1181
+ "num_tokens": 335380480.0,
1182
+ "step": 1280
1183
+ },
1184
+ {
1185
+ "epoch": 2.299598751671868,
1186
+ "grad_norm": 0.4882568120956421,
1187
+ "learning_rate": 3.0303640301979635e-05,
1188
+ "loss": 1.1982,
1189
+ "mean_token_accuracy": 0.7903318881988526,
1190
+ "num_tokens": 338001920.0,
1191
+ "step": 1290
1192
+ },
1193
+ {
1194
+ "epoch": 2.3174320106999553,
1195
+ "grad_norm": 0.5135630369186401,
1196
+ "learning_rate": 3.0014735394581823e-05,
1197
+ "loss": 1.2045,
1198
+ "mean_token_accuracy": 0.7893250770866871,
1199
+ "num_tokens": 340623360.0,
1200
+ "step": 1300
1201
+ },
1202
+ {
1203
+ "epoch": 2.3352652697280427,
1204
+ "grad_norm": 0.5496548414230347,
1205
+ "learning_rate": 2.972513100050851e-05,
1206
+ "loss": 1.1883,
1207
+ "mean_token_accuracy": 0.791755847632885,
1208
+ "num_tokens": 343244800.0,
1209
+ "step": 1310
1210
+ },
1211
+ {
1212
+ "epoch": 2.35309852875613,
1213
+ "grad_norm": 0.5601444840431213,
1214
+ "learning_rate": 2.943486751559295e-05,
1215
+ "loss": 1.1263,
1216
+ "mean_token_accuracy": 0.7998171858489513,
1217
+ "num_tokens": 345866240.0,
1218
+ "step": 1320
1219
+ },
1220
+ {
1221
+ "epoch": 2.3709317877842175,
1222
+ "grad_norm": 0.5243608355522156,
1223
+ "learning_rate": 2.9143985427602516e-05,
1224
+ "loss": 1.2274,
1225
+ "mean_token_accuracy": 0.7866500526666641,
1226
+ "num_tokens": 348487680.0,
1227
+ "step": 1330
1228
+ },
1229
+ {
1230
+ "epoch": 2.388765046812305,
1231
+ "grad_norm": 0.573567807674408,
1232
+ "learning_rate": 2.88525253105912e-05,
1233
+ "loss": 1.1977,
1234
+ "mean_token_accuracy": 0.7901960171759128,
1235
+ "num_tokens": 351109120.0,
1236
+ "step": 1340
1237
+ },
1238
+ {
1239
+ "epoch": 2.4065983058403924,
1240
+ "grad_norm": 0.4823901951313019,
1241
+ "learning_rate": 2.856052781924012e-05,
1242
+ "loss": 1.1358,
1243
+ "mean_token_accuracy": 0.7998057343065739,
1244
+ "num_tokens": 353730560.0,
1245
+ "step": 1350
1246
+ },
1247
+ {
1248
+ "epoch": 2.4244315648684798,
1249
+ "grad_norm": 0.515941858291626,
1250
+ "learning_rate": 2.8268033683186697e-05,
1251
+ "loss": 1.1881,
1252
+ "mean_token_accuracy": 0.7909879550337792,
1253
+ "num_tokens": 356352000.0,
1254
+ "step": 1360
1255
+ },
1256
+ {
1257
+ "epoch": 2.442264823896567,
1258
+ "grad_norm": 0.5508431196212769,
1259
+ "learning_rate": 2.7975083701343497e-05,
1260
+ "loss": 1.171,
1261
+ "mean_token_accuracy": 0.7945770494639873,
1262
+ "num_tokens": 358973440.0,
1263
+ "step": 1370
1264
+ },
1265
+ {
1266
+ "epoch": 2.4600980829246546,
1267
+ "grad_norm": 0.5473787188529968,
1268
+ "learning_rate": 2.7681718736207298e-05,
1269
+ "loss": 1.1309,
1270
+ "mean_token_accuracy": 0.7995454497635365,
1271
+ "num_tokens": 361594880.0,
1272
+ "step": 1380
1273
+ },
1274
+ {
1275
+ "epoch": 2.477931341952742,
1276
+ "grad_norm": 0.6680080890655518,
1277
+ "learning_rate": 2.738797970815933e-05,
1278
+ "loss": 1.1848,
1279
+ "mean_token_accuracy": 0.7917638584971428,
1280
+ "num_tokens": 364216320.0,
1281
+ "step": 1390
1282
+ },
1283
+ {
1284
+ "epoch": 2.4957646009808294,
1285
+ "grad_norm": 0.49958300590515137,
1286
+ "learning_rate": 2.7093907589757468e-05,
1287
+ "loss": 1.1922,
1288
+ "mean_token_accuracy": 0.7899307683110237,
1289
+ "num_tokens": 366837760.0,
1290
+ "step": 1400
1291
+ },
1292
+ {
1293
+ "epoch": 2.5135978600089164,
1294
+ "grad_norm": 0.4601416289806366,
1295
+ "learning_rate": 2.6799543400021132e-05,
1296
+ "loss": 1.1448,
1297
+ "mean_token_accuracy": 0.7981104202568531,
1298
+ "num_tokens": 369459200.0,
1299
+ "step": 1410
1300
+ },
1301
+ {
1302
+ "epoch": 2.531431119037004,
1303
+ "grad_norm": 0.5619862079620361,
1304
+ "learning_rate": 2.650492819870965e-05,
1305
+ "loss": 1.1371,
1306
+ "mean_token_accuracy": 0.7994878143072128,
1307
+ "num_tokens": 372080640.0,
1308
+ "step": 1420
1309
+ },
1310
+ {
1311
+ "epoch": 2.549264378065091,
1312
+ "grad_norm": 0.6030861139297485,
1313
+ "learning_rate": 2.621010308059504e-05,
1314
+ "loss": 1.1695,
1315
+ "mean_token_accuracy": 0.7944411806762218,
1316
+ "num_tokens": 374702080.0,
1317
+ "step": 1430
1318
+ },
1319
+ {
1320
+ "epoch": 2.5670976370931786,
1321
+ "grad_norm": 0.5237954258918762,
1322
+ "learning_rate": 2.5915109169729814e-05,
1323
+ "loss": 1.1372,
1324
+ "mean_token_accuracy": 0.7982218623161316,
1325
+ "num_tokens": 377323520.0,
1326
+ "step": 1440
1327
+ },
1328
+ {
1329
+ "epoch": 2.584930896121266,
1330
+ "grad_norm": 0.5144650936126709,
1331
+ "learning_rate": 2.5619987613710756e-05,
1332
+ "loss": 1.1896,
1333
+ "mean_token_accuracy": 0.7915455557405948,
1334
+ "num_tokens": 379944960.0,
1335
+ "step": 1450
1336
+ },
1337
+ {
1338
+ "epoch": 2.6027641551493534,
1339
+ "grad_norm": 0.5404797196388245,
1340
+ "learning_rate": 2.5324779577939394e-05,
1341
+ "loss": 1.1364,
1342
+ "mean_token_accuracy": 0.7987703062593937,
1343
+ "num_tokens": 382566400.0,
1344
+ "step": 1460
1345
+ },
1346
+ {
1347
+ "epoch": 2.620597414177441,
1348
+ "grad_norm": 0.4770708382129669,
1349
+ "learning_rate": 2.5029526239880015e-05,
1350
+ "loss": 1.1541,
1351
+ "mean_token_accuracy": 0.7973127588629723,
1352
+ "num_tokens": 385187840.0,
1353
+ "step": 1470
1354
+ },
1355
+ {
1356
+ "epoch": 2.6384306732055283,
1357
+ "grad_norm": 0.5579174757003784,
1358
+ "learning_rate": 2.473426878331591e-05,
1359
+ "loss": 1.1695,
1360
+ "mean_token_accuracy": 0.7945644512772561,
1361
+ "num_tokens": 387809280.0,
1362
+ "step": 1480
1363
+ },
1364
+ {
1365
+ "epoch": 2.6562639322336157,
1366
+ "grad_norm": 0.5287127494812012,
1367
+ "learning_rate": 2.443904839260488e-05,
1368
+ "loss": 1.1615,
1369
+ "mean_token_accuracy": 0.7949182473123073,
1370
+ "num_tokens": 390430720.0,
1371
+ "step": 1490
1372
+ },
1373
+ {
1374
+ "epoch": 2.674097191261703,
1375
+ "grad_norm": 0.5106226801872253,
1376
+ "learning_rate": 2.4143906246934537e-05,
1377
+ "loss": 1.1201,
1378
+ "mean_token_accuracy": 0.8007564425468445,
1379
+ "num_tokens": 393052160.0,
1380
+ "step": 1500
1381
+ },
1382
+ {
1383
+ "epoch": 2.6919304502897905,
1384
+ "grad_norm": 0.48907479643821716,
1385
+ "learning_rate": 2.384888351457838e-05,
1386
+ "loss": 1.1908,
1387
+ "mean_token_accuracy": 0.7911803111433983,
1388
+ "num_tokens": 395673600.0,
1389
+ "step": 1510
1390
+ },
1391
+ {
1392
+ "epoch": 2.709763709317878,
1393
+ "grad_norm": 0.5045855045318604,
1394
+ "learning_rate": 2.3554021347153403e-05,
1395
+ "loss": 1.1483,
1396
+ "mean_token_accuracy": 0.7968017213046551,
1397
+ "num_tokens": 398295040.0,
1398
+ "step": 1520
1399
+ },
1400
+ {
1401
+ "epoch": 2.7275969683459653,
1402
+ "grad_norm": 0.505662739276886,
1403
+ "learning_rate": 2.3259360873879992e-05,
1404
+ "loss": 1.1749,
1405
+ "mean_token_accuracy": 0.7930756144225597,
1406
+ "num_tokens": 400916480.0,
1407
+ "step": 1530
1408
+ },
1409
+ {
1410
+ "epoch": 2.7454302273740527,
1411
+ "grad_norm": 0.5283376574516296,
1412
+ "learning_rate": 2.2964943195844966e-05,
1413
+ "loss": 1.173,
1414
+ "mean_token_accuracy": 0.7921638391911984,
1415
+ "num_tokens": 403537920.0,
1416
+ "step": 1540
1417
+ },
1418
+ {
1419
+ "epoch": 2.76326348640214,
1420
+ "grad_norm": 0.5634790062904358,
1421
+ "learning_rate": 2.2670809380268548e-05,
1422
+ "loss": 1.1779,
1423
+ "mean_token_accuracy": 0.7936561174690724,
1424
+ "num_tokens": 406159360.0,
1425
+ "step": 1550
1426
+ },
1427
+ {
1428
+ "epoch": 2.7810967454302276,
1429
+ "grad_norm": 0.48228803277015686,
1430
+ "learning_rate": 2.237700045477607e-05,
1431
+ "loss": 1.14,
1432
+ "mean_token_accuracy": 0.796902097016573,
1433
+ "num_tokens": 408780800.0,
1434
+ "step": 1560
1435
+ },
1436
+ {
1437
+ "epoch": 2.798930004458315,
1438
+ "grad_norm": 0.4753395617008209,
1439
+ "learning_rate": 2.2083557401675115e-05,
1440
+ "loss": 1.1294,
1441
+ "mean_token_accuracy": 0.7987142026424408,
1442
+ "num_tokens": 411402240.0,
1443
+ "step": 1570
1444
+ },
1445
+ {
1446
+ "epoch": 2.8167632634864024,
1447
+ "grad_norm": 0.5348941683769226,
1448
+ "learning_rate": 2.1790521152239137e-05,
1449
+ "loss": 1.1581,
1450
+ "mean_token_accuracy": 0.7944865994155407,
1451
+ "num_tokens": 414023680.0,
1452
+ "step": 1580
1453
+ },
1454
+ {
1455
+ "epoch": 2.83459652251449,
1456
+ "grad_norm": 0.5610408782958984,
1457
+ "learning_rate": 2.1497932580998053e-05,
1458
+ "loss": 1.1609,
1459
+ "mean_token_accuracy": 0.795001070201397,
1460
+ "num_tokens": 416645120.0,
1461
+ "step": 1590
1462
+ },
1463
+ {
1464
+ "epoch": 2.8524297815425768,
1465
+ "grad_norm": 0.5431333184242249,
1466
+ "learning_rate": 2.1205832500036838e-05,
1467
+ "loss": 1.1405,
1468
+ "mean_token_accuracy": 0.7983539171516896,
1469
+ "num_tokens": 419266560.0,
1470
+ "step": 1600
1471
+ },
1472
+ {
1473
+ "epoch": 2.870263040570664,
1474
+ "grad_norm": 0.5650882720947266,
1475
+ "learning_rate": 2.0914261653302747e-05,
1476
+ "loss": 1.17,
1477
+ "mean_token_accuracy": 0.7925355695188046,
1478
+ "num_tokens": 421888000.0,
1479
+ "step": 1610
1480
+ },
1481
+ {
1482
+ "epoch": 2.8880962995987516,
1483
+ "grad_norm": 0.5435991287231445,
1484
+ "learning_rate": 2.062326071092218e-05,
1485
+ "loss": 1.1562,
1486
+ "mean_token_accuracy": 0.7965150982141495,
1487
+ "num_tokens": 424509440.0,
1488
+ "step": 1620
1489
+ },
1490
+ {
1491
+ "epoch": 2.905929558626839,
1492
+ "grad_norm": 0.4466145932674408,
1493
+ "learning_rate": 2.0332870263527675e-05,
1494
+ "loss": 1.1199,
1495
+ "mean_token_accuracy": 0.8016460791230202,
1496
+ "num_tokens": 427130880.0,
1497
+ "step": 1630
1498
+ },
1499
+ {
1500
+ "epoch": 2.9237628176549264,
1501
+ "grad_norm": 0.5083450675010681,
1502
+ "learning_rate": 2.0043130816596136e-05,
1503
+ "loss": 1.1467,
1504
+ "mean_token_accuracy": 0.7959990955889225,
1505
+ "num_tokens": 429752320.0,
1506
+ "step": 1640
1507
+ },
1508
+ {
1509
+ "epoch": 2.941596076683014,
1510
+ "grad_norm": 0.49085912108421326,
1511
+ "learning_rate": 1.9754082784798843e-05,
1512
+ "loss": 1.1565,
1513
+ "mean_token_accuracy": 0.7959273479878902,
1514
+ "num_tokens": 432373760.0,
1515
+ "step": 1650
1516
+ },
1517
+ {
1518
+ "epoch": 2.9594293357111012,
1519
+ "grad_norm": 0.48444733023643494,
1520
+ "learning_rate": 1.9465766486364143e-05,
1521
+ "loss": 1.1342,
1522
+ "mean_token_accuracy": 0.797958905249834,
1523
+ "num_tokens": 434995200.0,
1524
+ "step": 1660
1525
+ },
1526
+ {
1527
+ "epoch": 2.9772625947391886,
1528
+ "grad_norm": 0.5343450903892517,
1529
+ "learning_rate": 1.9178222137453656e-05,
1530
+ "loss": 1.141,
1531
+ "mean_token_accuracy": 0.7985142156481743,
1532
+ "num_tokens": 437616640.0,
1533
+ "step": 1670
1534
+ },
1535
+ {
1536
+ "epoch": 2.995095853767276,
1537
+ "grad_norm": 0.5583868026733398,
1538
+ "learning_rate": 1.8891489846552646e-05,
1539
+ "loss": 1.1432,
1540
+ "mean_token_accuracy": 0.7971971161663532,
1541
+ "num_tokens": 440238080.0,
1542
+ "step": 1680
1543
+ },
1544
+ {
1545
+ "epoch": 3.0,
1546
+ "eval_loss": 1.1234296560287476,
1547
+ "eval_mean_token_accuracy": 0.7894601821899414,
1548
+ "eval_num_tokens": 440942592.0,
1549
+ "eval_runtime": 2.9179,
1550
+ "eval_samples_per_second": 24.675,
1551
+ "eval_steps_per_second": 1.714,
1552
+ "step": 1683
1553
+ },
1554
+ {
1555
+ "epoch": 3.0124832813196614,
1556
+ "grad_norm": 0.5347409844398499,
1557
+ "learning_rate": 1.860560960887546e-05,
1558
+ "loss": 1.1253,
1559
+ "mean_token_accuracy": 0.8009015726737487,
1560
+ "num_tokens": 442777600.0,
1561
+ "step": 1690
1562
+ },
1563
+ {
1564
+ "epoch": 3.0303165403477483,
1565
+ "grad_norm": 0.5598998069763184,
1566
+ "learning_rate": 1.8320621300786688e-05,
1567
+ "loss": 1.106,
1568
+ "mean_token_accuracy": 0.8023651242256165,
1569
+ "num_tokens": 445399040.0,
1570
+ "step": 1700
1571
+ },
1572
+ {
1573
+ "epoch": 3.0481497993758357,
1574
+ "grad_norm": 0.5113139152526855,
1575
+ "learning_rate": 1.8036564674239043e-05,
1576
+ "loss": 1.0887,
1577
+ "mean_token_accuracy": 0.8040673099458218,
1578
+ "num_tokens": 448020480.0,
1579
+ "step": 1710
1580
+ },
1581
+ {
1582
+ "epoch": 3.065983058403923,
1583
+ "grad_norm": 0.44441869854927063,
1584
+ "learning_rate": 1.7753479351228444e-05,
1585
+ "loss": 1.1185,
1586
+ "mean_token_accuracy": 0.7999229028820991,
1587
+ "num_tokens": 450641920.0,
1588
+ "step": 1720
1589
+ },
1590
+ {
1591
+ "epoch": 3.0838163174320106,
1592
+ "grad_norm": 0.47339409589767456,
1593
+ "learning_rate": 1.7471404818267316e-05,
1594
+ "loss": 1.0925,
1595
+ "mean_token_accuracy": 0.8034070417284965,
1596
+ "num_tokens": 453263360.0,
1597
+ "step": 1730
1598
+ },
1599
+ {
1600
+ "epoch": 3.101649576460098,
1601
+ "grad_norm": 0.47856712341308594,
1602
+ "learning_rate": 1.7190380420876766e-05,
1603
+ "loss": 1.0911,
1604
+ "mean_token_accuracy": 0.8045188069343567,
1605
+ "num_tokens": 455884800.0,
1606
+ "step": 1740
1607
+ },
1608
+ {
1609
+ "epoch": 3.1194828354881854,
1610
+ "grad_norm": 0.4918275773525238,
1611
+ "learning_rate": 1.6910445358098415e-05,
1612
+ "loss": 1.1109,
1613
+ "mean_token_accuracy": 0.8015380732715129,
1614
+ "num_tokens": 458506240.0,
1615
+ "step": 1750
1616
+ },
1617
+ {
1618
+ "epoch": 3.137316094516273,
1619
+ "grad_norm": 0.5043804049491882,
1620
+ "learning_rate": 1.6631638677026674e-05,
1621
+ "loss": 1.1364,
1622
+ "mean_token_accuracy": 0.7974261090159416,
1623
+ "num_tokens": 461127680.0,
1624
+ "step": 1760
1625
+ },
1626
+ {
1627
+ "epoch": 3.15514935354436,
1628
+ "grad_norm": 0.4904772937297821,
1629
+ "learning_rate": 1.6353999267362237e-05,
1630
+ "loss": 1.0838,
1631
+ "mean_token_accuracy": 0.806386636942625,
1632
+ "num_tokens": 463749120.0,
1633
+ "step": 1770
1634
+ },
1635
+ {
1636
+ "epoch": 3.1729826125724476,
1637
+ "grad_norm": 0.48294153809547424,
1638
+ "learning_rate": 1.607756585598743e-05,
1639
+ "loss": 1.1389,
1640
+ "mean_token_accuracy": 0.7980314172804356,
1641
+ "num_tokens": 466370560.0,
1642
+ "step": 1780
1643
+ },
1644
+ {
1645
+ "epoch": 3.190815871600535,
1646
+ "grad_norm": 0.5139598250389099,
1647
+ "learning_rate": 1.580237700156443e-05,
1648
+ "loss": 1.1095,
1649
+ "mean_token_accuracy": 0.8025780878961086,
1650
+ "num_tokens": 468992000.0,
1651
+ "step": 1790
1652
+ },
1653
+ {
1654
+ "epoch": 3.2086491306286224,
1655
+ "grad_norm": 0.47298020124435425,
1656
+ "learning_rate": 1.5528471089156804e-05,
1657
+ "loss": 1.0931,
1658
+ "mean_token_accuracy": 0.8041585251688957,
1659
+ "num_tokens": 471613440.0,
1660
+ "step": 1800
1661
+ },
1662
+ {
1663
+ "epoch": 3.22648238965671,
1664
+ "grad_norm": 0.41903576254844666,
1665
+ "learning_rate": 1.5255886324875363e-05,
1666
+ "loss": 1.0954,
1667
+ "mean_token_accuracy": 0.8047371134161949,
1668
+ "num_tokens": 474234880.0,
1669
+ "step": 1810
1670
+ },
1671
+ {
1672
+ "epoch": 3.2443156486847973,
1673
+ "grad_norm": 0.4610045552253723,
1674
+ "learning_rate": 1.4984660730548856e-05,
1675
+ "loss": 1.1095,
1676
+ "mean_token_accuracy": 0.8022483348846435,
1677
+ "num_tokens": 476856320.0,
1678
+ "step": 1820
1679
+ },
1680
+ {
1681
+ "epoch": 3.2621489077128847,
1682
+ "grad_norm": 0.5105413198471069,
1683
+ "learning_rate": 1.4714832138420553e-05,
1684
+ "loss": 1.1258,
1685
+ "mean_token_accuracy": 0.7983997173607349,
1686
+ "num_tokens": 479477760.0,
1687
+ "step": 1830
1688
+ },
1689
+ {
1690
+ "epoch": 3.279982166740972,
1691
+ "grad_norm": 0.4393883943557739,
1692
+ "learning_rate": 1.4446438185871092e-05,
1693
+ "loss": 1.0566,
1694
+ "mean_token_accuracy": 0.8093036323785782,
1695
+ "num_tokens": 482099200.0,
1696
+ "step": 1840
1697
+ },
1698
+ {
1699
+ "epoch": 3.297815425769059,
1700
+ "grad_norm": 0.4618268311023712,
1701
+ "learning_rate": 1.4179516310168623e-05,
1702
+ "loss": 1.0697,
1703
+ "mean_token_accuracy": 0.8082189694046974,
1704
+ "num_tokens": 484720640.0,
1705
+ "step": 1850
1706
+ },
1707
+ {
1708
+ "epoch": 3.3156486847971465,
1709
+ "grad_norm": 0.47192856669425964,
1710
+ "learning_rate": 1.3914103743246757e-05,
1711
+ "loss": 1.1203,
1712
+ "mean_token_accuracy": 0.8002961635589599,
1713
+ "num_tokens": 487342080.0,
1714
+ "step": 1860
1715
+ },
1716
+ {
1717
+ "epoch": 3.333481943825234,
1718
+ "grad_norm": 0.4617849290370941,
1719
+ "learning_rate": 1.3650237506511331e-05,
1720
+ "loss": 1.0664,
1721
+ "mean_token_accuracy": 0.8089914366602897,
1722
+ "num_tokens": 489963520.0,
1723
+ "step": 1870
1724
+ },
1725
+ {
1726
+ "epoch": 3.3513152028533213,
1727
+ "grad_norm": 0.46263769268989563,
1728
+ "learning_rate": 1.3387954405676329e-05,
1729
+ "loss": 1.0855,
1730
+ "mean_token_accuracy": 0.8045512497425079,
1731
+ "num_tokens": 492584960.0,
1732
+ "step": 1880
1733
+ },
1734
+ {
1735
+ "epoch": 3.3691484618814087,
1736
+ "grad_norm": 0.47842440009117126,
1737
+ "learning_rate": 1.3127291025630073e-05,
1738
+ "loss": 1.0705,
1739
+ "mean_token_accuracy": 0.8085929870605468,
1740
+ "num_tokens": 495206400.0,
1741
+ "step": 1890
1742
+ },
1743
+ {
1744
+ "epoch": 3.386981720909496,
1745
+ "grad_norm": 0.43500861525535583,
1746
+ "learning_rate": 1.2868283725332028e-05,
1747
+ "loss": 1.0635,
1748
+ "mean_token_accuracy": 0.8091395162045956,
1749
+ "num_tokens": 497827840.0,
1750
+ "step": 1900
1751
+ },
1752
+ {
1753
+ "epoch": 3.4048149799375835,
1754
+ "grad_norm": 0.508797287940979,
1755
+ "learning_rate": 1.261096863274136e-05,
1756
+ "loss": 1.1052,
1757
+ "mean_token_accuracy": 0.8021300226449967,
1758
+ "num_tokens": 500449280.0,
1759
+ "step": 1910
1760
+ },
1761
+ {
1762
+ "epoch": 3.422648238965671,
1763
+ "grad_norm": 0.42378219962120056,
1764
+ "learning_rate": 1.2355381639777472e-05,
1765
+ "loss": 1.0936,
1766
+ "mean_token_accuracy": 0.8043497301638126,
1767
+ "num_tokens": 503070720.0,
1768
+ "step": 1920
1769
+ },
1770
+ {
1771
+ "epoch": 3.4404814979937584,
1772
+ "grad_norm": 0.4547331631183624,
1773
+ "learning_rate": 1.2101558397313634e-05,
1774
+ "loss": 1.1142,
1775
+ "mean_token_accuracy": 0.8015869230031967,
1776
+ "num_tokens": 505692160.0,
1777
+ "step": 1930
1778
+ },
1779
+ {
1780
+ "epoch": 3.4583147570218458,
1781
+ "grad_norm": 0.4319451153278351,
1782
+ "learning_rate": 1.1849534310204152e-05,
1783
+ "loss": 1.0835,
1784
+ "mean_token_accuracy": 0.8058400988578797,
1785
+ "num_tokens": 508313600.0,
1786
+ "step": 1940
1787
+ },
1788
+ {
1789
+ "epoch": 3.476148016049933,
1790
+ "grad_norm": 0.4666835069656372,
1791
+ "learning_rate": 1.1599344532345877e-05,
1792
+ "loss": 1.1072,
1793
+ "mean_token_accuracy": 0.801591893285513,
1794
+ "num_tokens": 510935040.0,
1795
+ "step": 1950
1796
+ },
1797
+ {
1798
+ "epoch": 3.4939812750780206,
1799
+ "grad_norm": 0.47252121567726135,
1800
+ "learning_rate": 1.1351023961774749e-05,
1801
+ "loss": 1.1245,
1802
+ "mean_token_accuracy": 0.7985180281102657,
1803
+ "num_tokens": 513556480.0,
1804
+ "step": 1960
1805
+ },
1806
+ {
1807
+ "epoch": 3.511814534106108,
1808
+ "grad_norm": 0.5152117013931274,
1809
+ "learning_rate": 1.110460723579796e-05,
1810
+ "loss": 1.0982,
1811
+ "mean_token_accuracy": 0.8033757455646992,
1812
+ "num_tokens": 516177920.0,
1813
+ "step": 1970
1814
+ },
1815
+ {
1816
+ "epoch": 3.5296477931341954,
1817
+ "grad_norm": 0.43357202410697937,
1818
+ "learning_rate": 1.0860128726162545e-05,
1819
+ "loss": 1.0632,
1820
+ "mean_token_accuracy": 0.8095081992447376,
1821
+ "num_tokens": 518799360.0,
1822
+ "step": 1980
1823
+ },
1824
+ {
1825
+ "epoch": 3.547481052162283,
1826
+ "grad_norm": 0.45368409156799316,
1827
+ "learning_rate": 1.0617622534260985e-05,
1828
+ "loss": 1.114,
1829
+ "mean_token_accuracy": 0.8014212884008884,
1830
+ "num_tokens": 521420800.0,
1831
+ "step": 1990
1832
+ },
1833
+ {
1834
+ "epoch": 3.5653143111903702,
1835
+ "grad_norm": 0.38939619064331055,
1836
+ "learning_rate": 1.0377122486374576e-05,
1837
+ "loss": 1.0449,
1838
+ "mean_token_accuracy": 0.8123023062944412,
1839
+ "num_tokens": 524042240.0,
1840
+ "step": 2000
1841
+ },
1842
+ {
1843
+ "epoch": 3.5831475702184576,
1844
+ "grad_norm": 0.42768579721450806,
1845
+ "learning_rate": 1.0138662128955053e-05,
1846
+ "loss": 1.104,
1847
+ "mean_token_accuracy": 0.8025334358215332,
1848
+ "num_tokens": 526663680.0,
1849
+ "step": 2010
1850
+ },
1851
+ {
1852
+ "epoch": 3.600980829246545,
1853
+ "grad_norm": 0.41337716579437256,
1854
+ "learning_rate": 9.902274723945381e-06,
1855
+ "loss": 1.0788,
1856
+ "mean_token_accuracy": 0.8076182380318642,
1857
+ "num_tokens": 529285120.0,
1858
+ "step": 2020
1859
+ },
1860
+ {
1861
+ "epoch": 3.618814088274632,
1862
+ "grad_norm": 0.4830559194087982,
1863
+ "learning_rate": 9.667993244140134e-06,
1864
+ "loss": 1.0967,
1865
+ "mean_token_accuracy": 0.8028597474098206,
1866
+ "num_tokens": 531906560.0,
1867
+ "step": 2030
1868
+ },
1869
+ {
1870
+ "epoch": 3.6366473473027194,
1871
+ "grad_norm": 0.4195765554904938,
1872
+ "learning_rate": 9.435850368586286e-06,
1873
+ "loss": 1.0773,
1874
+ "mean_token_accuracy": 0.8066488340497017,
1875
+ "num_tokens": 534528000.0,
1876
+ "step": 2040
1877
+ },
1878
+ {
1879
+ "epoch": 3.654480606330807,
1880
+ "grad_norm": 0.4251314103603363,
1881
+ "learning_rate": 9.20587847802488e-06,
1882
+ "loss": 1.0819,
1883
+ "mean_token_accuracy": 0.8052599862217903,
1884
+ "num_tokens": 537149440.0,
1885
+ "step": 2050
1886
+ },
1887
+ {
1888
+ "epoch": 3.6723138653588943,
1889
+ "grad_norm": 0.4114632308483124,
1890
+ "learning_rate": 8.978109650374397e-06,
1891
+ "loss": 1.0697,
1892
+ "mean_token_accuracy": 0.8077415138483047,
1893
+ "num_tokens": 539770880.0,
1894
+ "step": 2060
1895
+ },
1896
+ {
1897
+ "epoch": 3.6901471243869817,
1898
+ "grad_norm": 0.4304318428039551,
1899
+ "learning_rate": 8.752575656256321e-06,
1900
+ "loss": 1.1046,
1901
+ "mean_token_accuracy": 0.8016697481274605,
1902
+ "num_tokens": 542392320.0,
1903
+ "step": 2070
1904
+ },
1905
+ {
1906
+ "epoch": 3.707980383415069,
1907
+ "grad_norm": 0.47687673568725586,
1908
+ "learning_rate": 8.52930795456355e-06,
1909
+ "loss": 1.0558,
1910
+ "mean_token_accuracy": 0.8107806429266929,
1911
+ "num_tokens": 545013760.0,
1912
+ "step": 2080
1913
+ },
1914
+ {
1915
+ "epoch": 3.7258136424431565,
1916
+ "grad_norm": 0.45524051785469055,
1917
+ "learning_rate": 8.30833768807237e-06,
1918
+ "loss": 1.0816,
1919
+ "mean_token_accuracy": 0.8059305518865585,
1920
+ "num_tokens": 547635200.0,
1921
+ "step": 2090
1922
+ },
1923
+ {
1924
+ "epoch": 3.743646901471244,
1925
+ "grad_norm": 0.45478421449661255,
1926
+ "learning_rate": 8.089695679098375e-06,
1927
+ "loss": 1.0961,
1928
+ "mean_token_accuracy": 0.8029544025659561,
1929
+ "num_tokens": 550256640.0,
1930
+ "step": 2100
1931
+ },
1932
+ {
1933
+ "epoch": 3.7614801604993313,
1934
+ "grad_norm": 0.45910394191741943,
1935
+ "learning_rate": 7.87341242519728e-06,
1936
+ "loss": 1.0608,
1937
+ "mean_token_accuracy": 0.8082987323403359,
1938
+ "num_tokens": 552878080.0,
1939
+ "step": 2110
1940
+ },
1941
+ {
1942
+ "epoch": 3.7793134195274187,
1943
+ "grad_norm": 0.4360869824886322,
1944
+ "learning_rate": 7.659518094910839e-06,
1945
+ "loss": 1.0806,
1946
+ "mean_token_accuracy": 0.8052332706749439,
1947
+ "num_tokens": 555499520.0,
1948
+ "step": 2120
1949
+ },
1950
+ {
1951
+ "epoch": 3.797146678555506,
1952
+ "grad_norm": 0.46505963802337646,
1953
+ "learning_rate": 7.44804252355884e-06,
1954
+ "loss": 1.1143,
1955
+ "mean_token_accuracy": 0.8007854476571084,
1956
+ "num_tokens": 558120960.0,
1957
+ "step": 2130
1958
+ },
1959
+ {
1960
+ "epoch": 3.8149799375835935,
1961
+ "grad_norm": 0.40090060234069824,
1962
+ "learning_rate": 7.2390152090773765e-06,
1963
+ "loss": 1.079,
1964
+ "mean_token_accuracy": 0.8058068990707398,
1965
+ "num_tokens": 560742400.0,
1966
+ "step": 2140
1967
+ },
1968
+ {
1969
+ "epoch": 3.8328131966116805,
1970
+ "grad_norm": 0.41377872228622437,
1971
+ "learning_rate": 7.032465307904404e-06,
1972
+ "loss": 1.0496,
1973
+ "mean_token_accuracy": 0.8097444467246533,
1974
+ "num_tokens": 563363840.0,
1975
+ "step": 2150
1976
+ },
1977
+ {
1978
+ "epoch": 3.850646455639768,
1979
+ "grad_norm": 0.45236936211586,
1980
+ "learning_rate": 6.82842163091274e-06,
1981
+ "loss": 1.0819,
1982
+ "mean_token_accuracy": 0.8053412742912769,
1983
+ "num_tokens": 565985280.0,
1984
+ "step": 2160
1985
+ },
1986
+ {
1987
+ "epoch": 3.8684797146678553,
1988
+ "grad_norm": 0.42451992630958557,
1989
+ "learning_rate": 6.626912639391377e-06,
1990
+ "loss": 1.0649,
1991
+ "mean_token_accuracy": 0.8076308310031891,
1992
+ "num_tokens": 568606720.0,
1993
+ "step": 2170
1994
+ },
1995
+ {
1996
+ "epoch": 3.8863129736959428,
1997
+ "grad_norm": 0.43241724371910095,
1998
+ "learning_rate": 6.427966441075511e-06,
1999
+ "loss": 1.0821,
2000
+ "mean_token_accuracy": 0.805367611348629,
2001
+ "num_tokens": 571228160.0,
2002
+ "step": 2180
2003
+ },
2004
+ {
2005
+ "epoch": 3.90414623272403,
2006
+ "grad_norm": 0.41288912296295166,
2007
+ "learning_rate": 6.2316107862259026e-06,
2008
+ "loss": 1.1017,
2009
+ "mean_token_accuracy": 0.8022395603358745,
2010
+ "num_tokens": 573849600.0,
2011
+ "step": 2190
2012
+ },
2013
+ {
2014
+ "epoch": 3.9219794917521176,
2015
+ "grad_norm": 0.4336395263671875,
2016
+ "learning_rate": 6.037873063758112e-06,
2017
+ "loss": 1.0546,
2018
+ "mean_token_accuracy": 0.8097028434276581,
2019
+ "num_tokens": 576471040.0,
2020
+ "step": 2200
2021
+ },
2022
+ {
2023
+ "epoch": 3.939812750780205,
2024
+ "grad_norm": 0.4041893482208252,
2025
+ "learning_rate": 5.846780297422091e-06,
2026
+ "loss": 1.0603,
2027
+ "mean_token_accuracy": 0.8089009828865528,
2028
+ "num_tokens": 579092480.0,
2029
+ "step": 2210
2030
+ },
2031
+ {
2032
+ "epoch": 3.9576460098082924,
2033
+ "grad_norm": 0.39978188276290894,
2034
+ "learning_rate": 5.6583591420327684e-06,
2035
+ "loss": 1.0752,
2036
+ "mean_token_accuracy": 0.8069194242358207,
2037
+ "num_tokens": 581713920.0,
2038
+ "step": 2220
2039
+ },
2040
+ {
2041
+ "epoch": 3.97547926883638,
2042
+ "grad_norm": 0.41306570172309875,
2043
+ "learning_rate": 5.472635879752042e-06,
2044
+ "loss": 1.0757,
2045
+ "mean_token_accuracy": 0.8059477277100087,
2046
+ "num_tokens": 584335360.0,
2047
+ "step": 2230
2048
+ },
2049
+ {
2050
+ "epoch": 3.993312527864467,
2051
+ "grad_norm": 0.4413711428642273,
2052
+ "learning_rate": 5.289636416422822e-06,
2053
+ "loss": 1.1117,
2054
+ "mean_token_accuracy": 0.8009537614881992,
2055
+ "num_tokens": 586956800.0,
2056
+ "step": 2240
2057
+ },
2058
+ {
2059
+ "epoch": 4.0,
2060
+ "eval_loss": 1.0707859992980957,
2061
+ "eval_mean_token_accuracy": 0.7980337142944336,
2062
+ "eval_num_tokens": 587923456.0,
2063
+ "eval_runtime": 3.1887,
2064
+ "eval_samples_per_second": 22.579,
2065
+ "eval_steps_per_second": 1.568,
2066
+ "step": 2244
2067
+ },
2068
+ {
2069
+ "epoch": 4.010699955416852,
2070
+ "grad_norm": 0.3837050199508667,
2071
+ "learning_rate": 5.1093862779554776e-06,
2072
+ "loss": 1.0237,
2073
+ "mean_token_accuracy": 0.814745701276339,
2074
+ "num_tokens": 589496320.0,
2075
+ "step": 2250
2076
+ },
2077
+ {
2078
+ "epoch": 4.0285332144449395,
2079
+ "grad_norm": 0.39410483837127686,
2080
+ "learning_rate": 4.93191060676734e-06,
2081
+ "loss": 1.059,
2082
+ "mean_token_accuracy": 0.8092822574079037,
2083
+ "num_tokens": 592117760.0,
2084
+ "step": 2260
2085
+ },
2086
+ {
2087
+ "epoch": 4.046366473473027,
2088
+ "grad_norm": 0.4275108873844147,
2089
+ "learning_rate": 4.757234158275681e-06,
2090
+ "loss": 1.0876,
2091
+ "mean_token_accuracy": 0.8050832770764828,
2092
+ "num_tokens": 594739200.0,
2093
+ "step": 2270
2094
+ },
2095
+ {
2096
+ "epoch": 4.064199732501114,
2097
+ "grad_norm": 0.39579182863235474,
2098
+ "learning_rate": 4.585381297444688e-06,
2099
+ "loss": 1.0313,
2100
+ "mean_token_accuracy": 0.8126221284270286,
2101
+ "num_tokens": 597360640.0,
2102
+ "step": 2280
2103
+ },
2104
+ {
2105
+ "epoch": 4.082032991529202,
2106
+ "grad_norm": 0.4174662232398987,
2107
+ "learning_rate": 4.416375995386857e-06,
2108
+ "loss": 1.0677,
2109
+ "mean_token_accuracy": 0.808067063242197,
2110
+ "num_tokens": 599982080.0,
2111
+ "step": 2290
2112
+ },
2113
+ {
2114
+ "epoch": 4.099866250557289,
2115
+ "grad_norm": 0.426666796207428,
2116
+ "learning_rate": 4.2502418260193686e-06,
2117
+ "loss": 1.0551,
2118
+ "mean_token_accuracy": 0.8088761754333973,
2119
+ "num_tokens": 602603520.0,
2120
+ "step": 2300
2121
+ },
2122
+ {
2123
+ "epoch": 4.117699509585377,
2124
+ "grad_norm": 0.3763045370578766,
2125
+ "learning_rate": 4.087001962775849e-06,
2126
+ "loss": 1.0624,
2127
+ "mean_token_accuracy": 0.8099226720631123,
2128
+ "num_tokens": 605224960.0,
2129
+ "step": 2310
2130
+ },
2131
+ {
2132
+ "epoch": 4.135532768613464,
2133
+ "grad_norm": 0.40583643317222595,
2134
+ "learning_rate": 3.926679175373985e-06,
2135
+ "loss": 1.0593,
2136
+ "mean_token_accuracy": 0.8088261790573597,
2137
+ "num_tokens": 607846400.0,
2138
+ "step": 2320
2139
+ },
2140
+ {
2141
+ "epoch": 4.153366027641551,
2142
+ "grad_norm": 0.3805159330368042,
2143
+ "learning_rate": 3.7692958266394963e-06,
2144
+ "loss": 1.0686,
2145
+ "mean_token_accuracy": 0.808470857143402,
2146
+ "num_tokens": 610467840.0,
2147
+ "step": 2330
2148
+ },
2149
+ {
2150
+ "epoch": 4.171199286669639,
2151
+ "grad_norm": 0.36603036522865295,
2152
+ "learning_rate": 3.614873869386762e-06,
2153
+ "loss": 1.0475,
2154
+ "mean_token_accuracy": 0.8111096285283566,
2155
+ "num_tokens": 613089280.0,
2156
+ "step": 2340
2157
+ },
2158
+ {
2159
+ "epoch": 4.189032545697726,
2160
+ "grad_norm": 0.4105148911476135,
2161
+ "learning_rate": 3.463434843356783e-06,
2162
+ "loss": 1.0474,
2163
+ "mean_token_accuracy": 0.8102818146348,
2164
+ "num_tokens": 615710720.0,
2165
+ "step": 2350
2166
+ },
2167
+ {
2168
+ "epoch": 4.206865804725814,
2169
+ "grad_norm": 0.41211214661598206,
2170
+ "learning_rate": 3.314999872212618e-06,
2171
+ "loss": 1.1055,
2172
+ "mean_token_accuracy": 0.8027391448616982,
2173
+ "num_tokens": 618332160.0,
2174
+ "step": 2360
2175
+ },
2176
+ {
2177
+ "epoch": 4.224699063753901,
2178
+ "grad_norm": 0.409272164106369,
2179
+ "learning_rate": 3.169589660592978e-06,
2180
+ "loss": 1.0479,
2181
+ "mean_token_accuracy": 0.8102260909974575,
2182
+ "num_tokens": 620953600.0,
2183
+ "step": 2370
2184
+ },
2185
+ {
2186
+ "epoch": 4.242532322781988,
2187
+ "grad_norm": 0.4126708209514618,
2188
+ "learning_rate": 3.0272244912241453e-06,
2189
+ "loss": 1.0142,
2190
+ "mean_token_accuracy": 0.8164093054831028,
2191
+ "num_tokens": 623575040.0,
2192
+ "step": 2380
2193
+ },
2194
+ {
2195
+ "epoch": 4.260365581810076,
2196
+ "grad_norm": 0.3964686393737793,
2197
+ "learning_rate": 2.887924222090899e-06,
2198
+ "loss": 1.0644,
2199
+ "mean_token_accuracy": 0.8078506626188755,
2200
+ "num_tokens": 626196480.0,
2201
+ "step": 2390
2202
+ },
2203
+ {
2204
+ "epoch": 4.278198840838163,
2205
+ "grad_norm": 0.40106719732284546,
2206
+ "learning_rate": 2.751708283666521e-06,
2207
+ "loss": 1.0834,
2208
+ "mean_token_accuracy": 0.8057584255933762,
2209
+ "num_tokens": 628817920.0,
2210
+ "step": 2400
2211
+ },
2212
+ {
2213
+ "epoch": 4.296032099866251,
2214
+ "grad_norm": 0.3820619285106659,
2215
+ "learning_rate": 2.6185956762025592e-06,
2216
+ "loss": 1.0744,
2217
+ "mean_token_accuracy": 0.8067240163683891,
2218
+ "num_tokens": 631439360.0,
2219
+ "step": 2410
2220
+ },
2221
+ {
2222
+ "epoch": 4.313865358894338,
2223
+ "grad_norm": 0.41071221232414246,
2224
+ "learning_rate": 2.488604967078528e-06,
2225
+ "loss": 1.0805,
2226
+ "mean_token_accuracy": 0.8062465645372867,
2227
+ "num_tokens": 634060800.0,
2228
+ "step": 2420
2229
+ },
2230
+ {
2231
+ "epoch": 4.3316986179224255,
2232
+ "grad_norm": 0.3963005542755127,
2233
+ "learning_rate": 2.361754288212031e-06,
2234
+ "loss": 1.0898,
2235
+ "mean_token_accuracy": 0.8040848657488823,
2236
+ "num_tokens": 636682240.0,
2237
+ "step": 2430
2238
+ },
2239
+ {
2240
+ "epoch": 4.349531876950513,
2241
+ "grad_norm": 0.3937000036239624,
2242
+ "learning_rate": 2.2380613335296036e-06,
2243
+ "loss": 1.0211,
2244
+ "mean_token_accuracy": 0.8141208916902543,
2245
+ "num_tokens": 639303680.0,
2246
+ "step": 2440
2247
+ },
2248
+ {
2249
+ "epoch": 4.3673651359786,
2250
+ "grad_norm": 0.401140034198761,
2251
+ "learning_rate": 2.1175433564986443e-06,
2252
+ "loss": 1.0464,
2253
+ "mean_token_accuracy": 0.8111970283091068,
2254
+ "num_tokens": 641925120.0,
2255
+ "step": 2450
2256
+ },
2257
+ {
2258
+ "epoch": 4.385198395006688,
2259
+ "grad_norm": 0.39180195331573486,
2260
+ "learning_rate": 2.0002171677208026e-06,
2261
+ "loss": 1.0422,
2262
+ "mean_token_accuracy": 0.811572577059269,
2263
+ "num_tokens": 644546560.0,
2264
+ "step": 2460
2265
+ },
2266
+ {
2267
+ "epoch": 4.403031654034775,
2268
+ "grad_norm": 0.3858877122402191,
2269
+ "learning_rate": 1.8860991325871214e-06,
2270
+ "loss": 1.0625,
2271
+ "mean_token_accuracy": 0.8080498903989792,
2272
+ "num_tokens": 647168000.0,
2273
+ "step": 2470
2274
+ },
2275
+ {
2276
+ "epoch": 4.4208649130628626,
2277
+ "grad_norm": 0.38444533944129944,
2278
+ "learning_rate": 1.7752051689953087e-06,
2279
+ "loss": 1.0375,
2280
+ "mean_token_accuracy": 0.8128984488546849,
2281
+ "num_tokens": 649789440.0,
2282
+ "step": 2480
2283
+ },
2284
+ {
2285
+ "epoch": 4.43869817209095,
2286
+ "grad_norm": 0.3802605867385864,
2287
+ "learning_rate": 1.6675507451293899e-06,
2288
+ "loss": 1.0732,
2289
+ "mean_token_accuracy": 0.8079281389713288,
2290
+ "num_tokens": 652410880.0,
2291
+ "step": 2490
2292
+ },
2293
+ {
2294
+ "epoch": 4.456531431119037,
2295
+ "grad_norm": 0.39424973726272583,
2296
+ "learning_rate": 1.5631508773021165e-06,
2297
+ "loss": 1.0759,
2298
+ "mean_token_accuracy": 0.8063347272574901,
2299
+ "num_tokens": 655032320.0,
2300
+ "step": 2500
2301
+ },
2302
+ {
2303
+ "epoch": 4.474364690147125,
2304
+ "grad_norm": 0.4043511748313904,
2305
+ "learning_rate": 1.4620201278603945e-06,
2306
+ "loss": 1.0533,
2307
+ "mean_token_accuracy": 0.8101692236959934,
2308
+ "num_tokens": 657653760.0,
2309
+ "step": 2510
2310
+ },
2311
+ {
2312
+ "epoch": 4.492197949175212,
2313
+ "grad_norm": 0.4222579002380371,
2314
+ "learning_rate": 1.3641726031540431e-06,
2315
+ "loss": 1.0429,
2316
+ "mean_token_accuracy": 0.8107840739190578,
2317
+ "num_tokens": 660275200.0,
2318
+ "step": 2520
2319
+ }
2320
+ ],
2321
+ "logging_steps": 10,
2322
+ "max_steps": 2800,
2323
+ "num_input_tokens_seen": 0,
2324
+ "num_train_epochs": 5,
2325
+ "save_steps": 140,
2326
+ "stateful_callbacks": {
2327
+ "TrainerControl": {
2328
+ "args": {
2329
+ "should_epoch_stop": false,
2330
+ "should_evaluate": false,
2331
+ "should_log": false,
2332
+ "should_save": true,
2333
+ "should_training_stop": false
2334
+ },
2335
+ "attributes": {}
2336
+ }
2337
+ },
2338
+ "total_flos": 3.287571416664244e+18,
2339
+ "train_batch_size": 16,
2340
+ "trial_name": null,
2341
+ "trial_params": null
2342
+ }
checkpoint-2520/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2660/config.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_configs": [
3
+ {
4
+ "vocab_size": 49152,
5
+ "max_position_embeddings": 8192,
6
+ "hidden_size": 960,
7
+ "intermediate_size": 2560,
8
+ "num_hidden_layers": 32,
9
+ "num_attention_heads": 15,
10
+ "num_key_value_heads": 5,
11
+ "hidden_act": "silu",
12
+ "initializer_range": 0.02,
13
+ "rms_norm_eps": 1e-05,
14
+ "pretraining_tp": 1,
15
+ "use_cache": true,
16
+ "rope_theta": 100000,
17
+ "rope_scaling": null,
18
+ "attention_bias": false,
19
+ "attention_dropout": 0.0,
20
+ "mlp_bias": false,
21
+ "head_dim": 64,
22
+ "return_dict": true,
23
+ "output_hidden_states": false,
24
+ "output_attentions": false,
25
+ "torchscript": false,
26
+ "torch_dtype": "float32",
27
+ "use_bfloat16": false,
28
+ "tf_legacy_loss": false,
29
+ "pruned_heads": {},
30
+ "tie_word_embeddings": true,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "is_decoder": false,
34
+ "cross_attention_hidden_size": null,
35
+ "add_cross_attention": false,
36
+ "tie_encoder_decoder": false,
37
+ "max_length": 20,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "early_stopping": false,
41
+ "num_beams": 1,
42
+ "num_beam_groups": 1,
43
+ "diversity_penalty": 0.0,
44
+ "temperature": 1.0,
45
+ "top_k": 50,
46
+ "top_p": 1.0,
47
+ "typical_p": 1.0,
48
+ "repetition_penalty": 1.0,
49
+ "length_penalty": 1.0,
50
+ "no_repeat_ngram_size": 0,
51
+ "encoder_no_repeat_ngram_size": 0,
52
+ "bad_words_ids": null,
53
+ "num_return_sequences": 1,
54
+ "output_scores": false,
55
+ "return_dict_in_generate": false,
56
+ "forced_bos_token_id": null,
57
+ "forced_eos_token_id": null,
58
+ "remove_invalid_values": false,
59
+ "exponential_decay_length_penalty": null,
60
+ "suppress_tokens": null,
61
+ "begin_suppress_tokens": null,
62
+ "architectures": [
63
+ "LlamaForCausalLM"
64
+ ],
65
+ "finetuning_task": null,
66
+ "id2label": {
67
+ "0": "LABEL_0",
68
+ "1": "LABEL_1"
69
+ },
70
+ "label2id": {
71
+ "LABEL_0": 0,
72
+ "LABEL_1": 1
73
+ },
74
+ "tokenizer_class": null,
75
+ "prefix": null,
76
+ "bos_token_id": 1,
77
+ "pad_token_id": 2,
78
+ "eos_token_id": 2,
79
+ "sep_token_id": null,
80
+ "decoder_start_token_id": null,
81
+ "task_specific_params": null,
82
+ "problem_type": null,
83
+ "_name_or_path": "HuggingFaceTB/SmolLM2-360M-Instruct",
84
+ "_attn_implementation_autoset": true,
85
+ "transformers_version": "4.51.3",
86
+ "is_llama_config": true,
87
+ "model_type": "llama",
88
+ "rope_interleaved": false,
89
+ "transformers.js_config": {
90
+ "kv_cache_dtype": {
91
+ "q4f16": "float16",
92
+ "fp16": "float16"
93
+ }
94
+ }
95
+ },
96
+ {
97
+ "vocab_size": 49152,
98
+ "max_position_embeddings": 8192,
99
+ "hidden_size": 576,
100
+ "intermediate_size": 1536,
101
+ "num_hidden_layers": 30,
102
+ "num_attention_heads": 9,
103
+ "num_key_value_heads": 3,
104
+ "hidden_act": "silu",
105
+ "initializer_range": 0.041666666666666664,
106
+ "rms_norm_eps": 1e-05,
107
+ "pretraining_tp": 1,
108
+ "use_cache": true,
109
+ "rope_theta": 100000,
110
+ "rope_scaling": null,
111
+ "attention_bias": false,
112
+ "attention_dropout": 0.0,
113
+ "mlp_bias": false,
114
+ "head_dim": 64,
115
+ "return_dict": true,
116
+ "output_hidden_states": false,
117
+ "output_attentions": false,
118
+ "torchscript": false,
119
+ "torch_dtype": "float32",
120
+ "use_bfloat16": false,
121
+ "tf_legacy_loss": false,
122
+ "pruned_heads": {},
123
+ "tie_word_embeddings": true,
124
+ "chunk_size_feed_forward": 0,
125
+ "is_encoder_decoder": false,
126
+ "is_decoder": false,
127
+ "cross_attention_hidden_size": null,
128
+ "add_cross_attention": false,
129
+ "tie_encoder_decoder": false,
130
+ "max_length": 20,
131
+ "min_length": 0,
132
+ "do_sample": false,
133
+ "early_stopping": false,
134
+ "num_beams": 1,
135
+ "num_beam_groups": 1,
136
+ "diversity_penalty": 0.0,
137
+ "temperature": 1.0,
138
+ "top_k": 50,
139
+ "top_p": 1.0,
140
+ "typical_p": 1.0,
141
+ "repetition_penalty": 1.0,
142
+ "length_penalty": 1.0,
143
+ "no_repeat_ngram_size": 0,
144
+ "encoder_no_repeat_ngram_size": 0,
145
+ "bad_words_ids": null,
146
+ "num_return_sequences": 1,
147
+ "output_scores": false,
148
+ "return_dict_in_generate": false,
149
+ "forced_bos_token_id": null,
150
+ "forced_eos_token_id": null,
151
+ "remove_invalid_values": false,
152
+ "exponential_decay_length_penalty": null,
153
+ "suppress_tokens": null,
154
+ "begin_suppress_tokens": null,
155
+ "architectures": [
156
+ "LlamaForCausalLM"
157
+ ],
158
+ "finetuning_task": null,
159
+ "id2label": {
160
+ "0": "LABEL_0",
161
+ "1": "LABEL_1"
162
+ },
163
+ "label2id": {
164
+ "LABEL_0": 0,
165
+ "LABEL_1": 1
166
+ },
167
+ "tokenizer_class": null,
168
+ "prefix": null,
169
+ "bos_token_id": 1,
170
+ "pad_token_id": 2,
171
+ "eos_token_id": 2,
172
+ "sep_token_id": null,
173
+ "decoder_start_token_id": null,
174
+ "task_specific_params": null,
175
+ "problem_type": null,
176
+ "_name_or_path": "HuggingFaceTB/SmolLM2-135M-Instruct",
177
+ "_attn_implementation_autoset": true,
178
+ "transformers_version": "4.51.3",
179
+ "is_llama_config": true,
180
+ "model_type": "llama",
181
+ "rope_interleaved": false,
182
+ "transformers.js_config": {
183
+ "kv_cache_dtype": {
184
+ "q4f16": "float16",
185
+ "fp16": "float16"
186
+ }
187
+ }
188
+ }
189
+ ],
190
+ "num_models": 2
191
+ }
checkpoint-2660/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2660/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }