Hugging Face's logo

roonbug
/
ep1ux3vt

Safetensors
Model card Files
xet
 Community
roonbug commited on
Commit
36bb2b4
·
verified ·
1 Parent(s): 4a6b559

Upload folder using huggingface_hub

Browse files
Files changed (20) hide show
  1. .gitattributes +1 -0
  2. checkpoint-500/added_tokens.json +3 -0
  3. checkpoint-500/chat_template.jinja +47 -0
  4. checkpoint-500/config.json +110 -0
  5. checkpoint-500/generation_config.json +13 -0
  6. checkpoint-500/model-00001-of-00005.safetensors +3 -0
  7. checkpoint-500/model-00002-of-00005.safetensors +3 -0
  8. checkpoint-500/model-00003-of-00005.safetensors +3 -0
  9. checkpoint-500/model-00004-of-00005.safetensors +3 -0
  10. checkpoint-500/model-00005-of-00005.safetensors +3 -0
  11. checkpoint-500/model.safetensors.index.json +0 -0
  12. checkpoint-500/optimizer.pt +3 -0
  13. checkpoint-500/rng_state.pth +3 -0
  14. checkpoint-500/scheduler.pt +3 -0
  15. checkpoint-500/special_tokens_map.json +33 -0
  16. checkpoint-500/tokenizer.json +3 -0
  17. checkpoint-500/tokenizer.model +3 -0
  18. checkpoint-500/tokenizer_config.json +0 -0
  19. checkpoint-500/trainer_state.json +644 -0
  20. checkpoint-500/training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
checkpoint-500/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<image_soft_token>": 262144
3
+ }
checkpoint-500/chat_template.jinja ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {{ bos_token }}
2
+ {%- if messages[0]['role'] == 'system' -%}
3
+ {%- if messages[0]['content'] is string -%}
4
+ {%- set first_user_prefix = messages[0]['content'] + '
5
+
6
+ ' -%}
7
+ {%- else -%}
8
+ {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
9
+
10
+ ' -%}
11
+ {%- endif -%}
12
+ {%- set loop_messages = messages[1:] -%}
13
+ {%- else -%}
14
+ {%- set first_user_prefix = "" -%}
15
+ {%- set loop_messages = messages -%}
16
+ {%- endif -%}
17
+ {%- for message in loop_messages -%}
18
+ {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
19
+ {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
20
+ {%- endif -%}
21
+ {%- if (message['role'] == 'assistant') -%}
22
+ {%- set role = "model" -%}
23
+ {%- else -%}
24
+ {%- set role = message['role'] -%}
25
+ {%- endif -%}
26
+ {{ '<start_of_turn>' + role + '
27
+ ' + (first_user_prefix if loop.first else "") }}
28
+ {%- if message['content'] is string -%}
29
+ {{ message['content'] | trim }}
30
+ {%- elif message['content'] is iterable -%}
31
+ {%- for item in message['content'] -%}
32
+ {%- if item['type'] == 'image' -%}
33
+ {{ '<start_of_image>' }}
34
+ {%- elif item['type'] == 'text' -%}
35
+ {{ item['text'] | trim }}
36
+ {%- endif -%}
37
+ {%- endfor -%}
38
+ {%- else -%}
39
+ {{ raise_exception("Invalid content type") }}
40
+ {%- endif -%}
41
+ {{ '<end_of_turn>
42
+ ' }}
43
+ {%- endfor -%}
44
+ {%- if add_generation_prompt -%}
45
+ {{'<start_of_turn>model
46
+ '}}
47
+ {%- endif -%}
checkpoint-500/config.json ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma3ForConditionalGeneration"
4
+ ],
5
+ "boi_token_index": 255999,
6
+ "bos_token_id": 2,
7
+ "dtype": "bfloat16",
8
+ "eoi_token_index": 256000,
9
+ "eos_token_id": 1,
10
+ "image_token_index": 262144,
11
+ "initializer_range": 0.02,
12
+ "mm_tokens_per_image": 256,
13
+ "model_type": "gemma3",
14
+ "pad_token_id": 0,
15
+ "text_config": {
16
+ "_sliding_window_pattern": 6,
17
+ "attention_bias": false,
18
+ "attention_dropout": 0.0,
19
+ "attn_logit_softcapping": null,
20
+ "dtype": "bfloat16",
21
+ "final_logit_softcapping": null,
22
+ "head_dim": 256,
23
+ "hidden_activation": "gelu_pytorch_tanh",
24
+ "hidden_size": 3840,
25
+ "initializer_range": 0.02,
26
+ "intermediate_size": 15360,
27
+ "layer_types": [
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "sliding_attention",
31
+ "sliding_attention",
32
+ "sliding_attention",
33
+ "full_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "sliding_attention",
38
+ "sliding_attention",
39
+ "full_attention",
40
+ "sliding_attention",
41
+ "sliding_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "sliding_attention",
45
+ "full_attention",
46
+ "sliding_attention",
47
+ "sliding_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "sliding_attention",
51
+ "full_attention",
52
+ "sliding_attention",
53
+ "sliding_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "sliding_attention",
57
+ "full_attention",
58
+ "sliding_attention",
59
+ "sliding_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "sliding_attention",
63
+ "full_attention",
64
+ "sliding_attention",
65
+ "sliding_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "sliding_attention",
69
+ "full_attention",
70
+ "sliding_attention",
71
+ "sliding_attention",
72
+ "sliding_attention",
73
+ "sliding_attention",
74
+ "sliding_attention",
75
+ "full_attention"
76
+ ],
77
+ "max_position_embeddings": 131072,
78
+ "model_type": "gemma3_text",
79
+ "num_attention_heads": 16,
80
+ "num_hidden_layers": 48,
81
+ "num_key_value_heads": 8,
82
+ "query_pre_attn_scalar": 256,
83
+ "rms_norm_eps": 1e-06,
84
+ "rope_local_base_freq": 10000.0,
85
+ "rope_scaling": {
86
+ "factor": 8.0,
87
+ "rope_type": "linear"
88
+ },
89
+ "rope_theta": 1000000.0,
90
+ "sliding_window": 1024,
91
+ "use_cache": true,
92
+ "vocab_size": 262208
93
+ },
94
+ "transformers_version": "4.56.1",
95
+ "vision_config": {
96
+ "attention_dropout": 0.0,
97
+ "dtype": "bfloat16",
98
+ "hidden_act": "gelu_pytorch_tanh",
99
+ "hidden_size": 1152,
100
+ "image_size": 896,
101
+ "intermediate_size": 4304,
102
+ "layer_norm_eps": 1e-06,
103
+ "model_type": "siglip_vision_model",
104
+ "num_attention_heads": 16,
105
+ "num_channels": 3,
106
+ "num_hidden_layers": 27,
107
+ "patch_size": 14,
108
+ "vision_use_head": false
109
+ }
110
+ }
checkpoint-500/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 2,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 1,
6
+ 1,
7
+ 106
8
+ ],
9
+ "pad_token_id": 0,
10
+ "top_k": 64,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.56.1"
13
+ }
checkpoint-500/model-00001-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4847447e92599833e8dbaa3067cd201c3bb5c052efa91f11ba891e43234f7832
3
+ size 4979902192
checkpoint-500/model-00002-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:891bd54eed03cba9ee1e705533a02a8217fcc29f356e4a1f53e5fd0d178883ad
3
+ size 4931296592
checkpoint-500/model-00003-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cee411d9d57324e50ce064a192cc5a858276d508611b12fc599e0c9767112e0
3
+ size 4931296656
checkpoint-500/model-00004-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fef54b57299c4de69bd369dc0f32c66d7edf3f113f8ec3cafb89861b8f9d464c
3
+ size 4931296656
checkpoint-500/model-00005-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66dc37f899acf501535c2a2afbdd9954a2e2ed6a5e163cd6944d830754676f6f
3
+ size 4601000928
checkpoint-500/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85dff2c1eef5e3f637bff445300e9c1e46777ddf7917fcce17278bacee009aa1
3
+ size 7286558019
checkpoint-500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69b6509f8bbdd943c9f8984cdc19d5d17ca63b6e47cd6e28080ebf502da06ddc
3
+ size 14645
checkpoint-500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65017fffec66ae2705d3217f0fed52569720fa202fff4a9be2acd26251c95e26
3
+ size 1465
checkpoint-500/special_tokens_map.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "boi_token": "<start_of_image>",
3
+ "bos_token": {
4
+ "content": "<bos>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eoi_token": "<end_of_image>",
11
+ "eos_token": {
12
+ "content": "<eos>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "image_token": "<image_soft_token>",
19
+ "pad_token": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false
25
+ },
26
+ "unk_token": {
27
+ "content": "<unk>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ }
33
+ }
checkpoint-500/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
3
+ size 33384568
checkpoint-500/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
3
+ size 4689074
checkpoint-500/tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-500/trainer_state.json ADDED
@@ -0,0 +1,644 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.8,
6
+ "eval_steps": 100,
7
+ "global_step": 500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "entropy": 0.7168930960819125,
14
+ "epoch": 0.016,
15
+ "grad_norm": 73.5,
16
+ "learning_rate": 1.8e-07,
17
+ "loss": 15.0906,
18
+ "mean_token_accuracy": 0.7680471498519182,
19
+ "num_tokens": 280941.0,
20
+ "step": 10
21
+ },
22
+ {
23
+ "entropy": 0.741888347826898,
24
+ "epoch": 0.032,
25
+ "grad_norm": 78.0,
26
+ "learning_rate": 3.8e-07,
27
+ "loss": 15.4631,
28
+ "mean_token_accuracy": 0.7632880255579948,
29
+ "num_tokens": 558056.0,
30
+ "step": 20
31
+ },
32
+ {
33
+ "entropy": 0.7365243999287486,
34
+ "epoch": 0.048,
35
+ "grad_norm": 65.0,
36
+ "learning_rate": 5.800000000000001e-07,
37
+ "loss": 15.1017,
38
+ "mean_token_accuracy": 0.768239913508296,
39
+ "num_tokens": 836753.0,
40
+ "step": 30
41
+ },
42
+ {
43
+ "entropy": 0.7818019269034266,
44
+ "epoch": 0.064,
45
+ "grad_norm": 56.25,
46
+ "learning_rate": 7.8e-07,
47
+ "loss": 15.6995,
48
+ "mean_token_accuracy": 0.7602430328726768,
49
+ "num_tokens": 1126446.0,
50
+ "step": 40
51
+ },
52
+ {
53
+ "entropy": 0.8105136282742024,
54
+ "epoch": 0.08,
55
+ "grad_norm": 47.25,
56
+ "learning_rate": 9.800000000000001e-07,
57
+ "loss": 15.0321,
58
+ "mean_token_accuracy": 0.7664048440754414,
59
+ "num_tokens": 1413596.0,
60
+ "step": 50
61
+ },
62
+ {
63
+ "entropy": 0.8216251201927662,
64
+ "epoch": 0.096,
65
+ "grad_norm": 39.75,
66
+ "learning_rate": 1.1800000000000001e-06,
67
+ "loss": 14.6285,
68
+ "mean_token_accuracy": 0.7707466218620539,
69
+ "num_tokens": 1701193.0,
70
+ "step": 60
71
+ },
72
+ {
73
+ "entropy": 0.8549969043582678,
74
+ "epoch": 0.112,
75
+ "grad_norm": 39.0,
76
+ "learning_rate": 1.3800000000000001e-06,
77
+ "loss": 14.8679,
78
+ "mean_token_accuracy": 0.7673181220889091,
79
+ "num_tokens": 1979232.0,
80
+ "step": 70
81
+ },
82
+ {
83
+ "entropy": 0.8556341353803874,
84
+ "epoch": 0.128,
85
+ "grad_norm": 34.0,
86
+ "learning_rate": 1.5800000000000001e-06,
87
+ "loss": 14.6467,
88
+ "mean_token_accuracy": 0.7710235182195901,
89
+ "num_tokens": 2274177.0,
90
+ "step": 80
91
+ },
92
+ {
93
+ "entropy": 0.8478256281465292,
94
+ "epoch": 0.144,
95
+ "grad_norm": 31.0,
96
+ "learning_rate": 1.7800000000000001e-06,
97
+ "loss": 14.3261,
98
+ "mean_token_accuracy": 0.7732654966413974,
99
+ "num_tokens": 2548445.0,
100
+ "step": 90
101
+ },
102
+ {
103
+ "entropy": 0.8438362725079059,
104
+ "epoch": 0.16,
105
+ "grad_norm": 38.5,
106
+ "learning_rate": 1.98e-06,
107
+ "loss": 14.0318,
108
+ "mean_token_accuracy": 0.7773757755756379,
109
+ "num_tokens": 2824418.0,
110
+ "step": 100
111
+ },
112
+ {
113
+ "epoch": 0.16,
114
+ "eval_biology_entropy": 1.211377203464508,
115
+ "eval_biology_loss": 1.1644827127456665,
116
+ "eval_biology_mean_token_accuracy": 0.7046201548576355,
117
+ "eval_biology_num_tokens": 2824418.0,
118
+ "eval_biology_runtime": 20.5128,
119
+ "eval_biology_samples_per_second": 24.375,
120
+ "eval_biology_steps_per_second": 6.094,
121
+ "step": 100
122
+ },
123
+ {
124
+ "epoch": 0.16,
125
+ "eval_math_entropy": 0.875089626789093,
126
+ "eval_math_loss": 0.8965557217597961,
127
+ "eval_math_mean_token_accuracy": 0.7736486663818359,
128
+ "eval_math_num_tokens": 2824418.0,
129
+ "eval_math_runtime": 25.0963,
130
+ "eval_math_samples_per_second": 19.923,
131
+ "eval_math_steps_per_second": 4.981,
132
+ "step": 100
133
+ },
134
+ {
135
+ "entropy": 0.8555477414280176,
136
+ "epoch": 0.176,
137
+ "grad_norm": 47.75,
138
+ "learning_rate": 2.1800000000000003e-06,
139
+ "loss": 14.0356,
140
+ "mean_token_accuracy": 0.7775060940533877,
141
+ "num_tokens": 3110313.0,
142
+ "step": 110
143
+ },
144
+ {
145
+ "entropy": 0.8610258311033249,
146
+ "epoch": 0.192,
147
+ "grad_norm": 31.375,
148
+ "learning_rate": 2.38e-06,
149
+ "loss": 14.025,
150
+ "mean_token_accuracy": 0.7762446004897356,
151
+ "num_tokens": 3394170.0,
152
+ "step": 120
153
+ },
154
+ {
155
+ "entropy": 0.8509209487587214,
156
+ "epoch": 0.208,
157
+ "grad_norm": 30.0,
158
+ "learning_rate": 2.5800000000000003e-06,
159
+ "loss": 13.8558,
160
+ "mean_token_accuracy": 0.7793015491217374,
161
+ "num_tokens": 3673600.0,
162
+ "step": 130
163
+ },
164
+ {
165
+ "entropy": 0.8540813602507115,
166
+ "epoch": 0.224,
167
+ "grad_norm": 30.125,
168
+ "learning_rate": 2.7800000000000005e-06,
169
+ "loss": 13.8247,
170
+ "mean_token_accuracy": 0.7789989039301872,
171
+ "num_tokens": 3953732.0,
172
+ "step": 140
173
+ },
174
+ {
175
+ "entropy": 0.8311325689777732,
176
+ "epoch": 0.24,
177
+ "grad_norm": 30.25,
178
+ "learning_rate": 2.9800000000000003e-06,
179
+ "loss": 13.4688,
180
+ "mean_token_accuracy": 0.7846441507339478,
181
+ "num_tokens": 4243655.0,
182
+ "step": 150
183
+ },
184
+ {
185
+ "entropy": 0.8143093746155501,
186
+ "epoch": 0.256,
187
+ "grad_norm": 27.625,
188
+ "learning_rate": 3.1800000000000005e-06,
189
+ "loss": 13.0712,
190
+ "mean_token_accuracy": 0.7881167802959681,
191
+ "num_tokens": 4531471.0,
192
+ "step": 160
193
+ },
194
+ {
195
+ "entropy": 0.8313567344099283,
196
+ "epoch": 0.272,
197
+ "grad_norm": 24.0,
198
+ "learning_rate": 3.3800000000000007e-06,
199
+ "loss": 13.3645,
200
+ "mean_token_accuracy": 0.7852793108671904,
201
+ "num_tokens": 4810284.0,
202
+ "step": 170
203
+ },
204
+ {
205
+ "entropy": 0.8148340426385403,
206
+ "epoch": 0.288,
207
+ "grad_norm": 26.375,
208
+ "learning_rate": 3.58e-06,
209
+ "loss": 13.0548,
210
+ "mean_token_accuracy": 0.7882425185292959,
211
+ "num_tokens": 5095104.0,
212
+ "step": 180
213
+ },
214
+ {
215
+ "entropy": 0.8263534324243664,
216
+ "epoch": 0.304,
217
+ "grad_norm": 25.125,
218
+ "learning_rate": 3.7800000000000002e-06,
219
+ "loss": 13.2184,
220
+ "mean_token_accuracy": 0.7860081434249878,
221
+ "num_tokens": 5383732.0,
222
+ "step": 190
223
+ },
224
+ {
225
+ "entropy": 0.7958642322570085,
226
+ "epoch": 0.32,
227
+ "grad_norm": 23.75,
228
+ "learning_rate": 3.980000000000001e-06,
229
+ "loss": 12.7072,
230
+ "mean_token_accuracy": 0.7937722463160753,
231
+ "num_tokens": 5676334.0,
232
+ "step": 200
233
+ },
234
+ {
235
+ "epoch": 0.32,
236
+ "eval_biology_entropy": 1.1982407326698303,
237
+ "eval_biology_loss": 1.1808913946151733,
238
+ "eval_biology_mean_token_accuracy": 0.6998598065376281,
239
+ "eval_biology_num_tokens": 5676334.0,
240
+ "eval_biology_runtime": 19.7103,
241
+ "eval_biology_samples_per_second": 25.367,
242
+ "eval_biology_steps_per_second": 6.342,
243
+ "step": 200
244
+ },
245
+ {
246
+ "epoch": 0.32,
247
+ "eval_math_entropy": 0.8109114770889282,
248
+ "eval_math_loss": 0.822318971157074,
249
+ "eval_math_mean_token_accuracy": 0.7874419956207276,
250
+ "eval_math_num_tokens": 5676334.0,
251
+ "eval_math_runtime": 24.5627,
252
+ "eval_math_samples_per_second": 20.356,
253
+ "eval_math_steps_per_second": 5.089,
254
+ "step": 200
255
+ },
256
+ {
257
+ "entropy": 0.8209991015493869,
258
+ "epoch": 0.336,
259
+ "grad_norm": 23.375,
260
+ "learning_rate": 4.18e-06,
261
+ "loss": 13.1713,
262
+ "mean_token_accuracy": 0.78699039965868,
263
+ "num_tokens": 5958480.0,
264
+ "step": 210
265
+ },
266
+ {
267
+ "entropy": 0.7989038350060582,
268
+ "epoch": 0.352,
269
+ "grad_norm": 25.0,
270
+ "learning_rate": 4.38e-06,
271
+ "loss": 12.7482,
272
+ "mean_token_accuracy": 0.7930307753384114,
273
+ "num_tokens": 6242161.0,
274
+ "step": 220
275
+ },
276
+ {
277
+ "entropy": 0.7915343299508095,
278
+ "epoch": 0.368,
279
+ "grad_norm": 24.875,
280
+ "learning_rate": 4.58e-06,
281
+ "loss": 12.6757,
282
+ "mean_token_accuracy": 0.7925275303423405,
283
+ "num_tokens": 6523679.0,
284
+ "step": 230
285
+ },
286
+ {
287
+ "entropy": 0.7743825454264879,
288
+ "epoch": 0.384,
289
+ "grad_norm": 23.375,
290
+ "learning_rate": 4.78e-06,
291
+ "loss": 12.3704,
292
+ "mean_token_accuracy": 0.79692403934896,
293
+ "num_tokens": 6810978.0,
294
+ "step": 240
295
+ },
296
+ {
297
+ "entropy": 0.8141555316746235,
298
+ "epoch": 0.4,
299
+ "grad_norm": 24.125,
300
+ "learning_rate": 4.980000000000001e-06,
301
+ "loss": 13.0618,
302
+ "mean_token_accuracy": 0.7885617177933455,
303
+ "num_tokens": 7096903.0,
304
+ "step": 250
305
+ },
306
+ {
307
+ "entropy": 0.7800503006204963,
308
+ "epoch": 0.416,
309
+ "grad_norm": 22.25,
310
+ "learning_rate": 5.18e-06,
311
+ "loss": 12.498,
312
+ "mean_token_accuracy": 0.7941101636737585,
313
+ "num_tokens": 7377181.0,
314
+ "step": 260
315
+ },
316
+ {
317
+ "entropy": 0.7713520534336566,
318
+ "epoch": 0.432,
319
+ "grad_norm": 25.625,
320
+ "learning_rate": 5.380000000000001e-06,
321
+ "loss": 12.2429,
322
+ "mean_token_accuracy": 0.798516258224845,
323
+ "num_tokens": 7650523.0,
324
+ "step": 270
325
+ },
326
+ {
327
+ "entropy": 0.767449575662613,
328
+ "epoch": 0.448,
329
+ "grad_norm": 24.875,
330
+ "learning_rate": 5.580000000000001e-06,
331
+ "loss": 12.2843,
332
+ "mean_token_accuracy": 0.7973004225641489,
333
+ "num_tokens": 7936788.0,
334
+ "step": 280
335
+ },
336
+ {
337
+ "entropy": 0.7630951976403594,
338
+ "epoch": 0.464,
339
+ "grad_norm": 21.375,
340
+ "learning_rate": 5.78e-06,
341
+ "loss": 12.2689,
342
+ "mean_token_accuracy": 0.7984702557325363,
343
+ "num_tokens": 8223147.0,
344
+ "step": 290
345
+ },
346
+ {
347
+ "entropy": 0.7839587140828371,
348
+ "epoch": 0.48,
349
+ "grad_norm": 22.125,
350
+ "learning_rate": 5.98e-06,
351
+ "loss": 12.4483,
352
+ "mean_token_accuracy": 0.7962926685810089,
353
+ "num_tokens": 8506030.0,
354
+ "step": 300
355
+ },
356
+ {
357
+ "epoch": 0.48,
358
+ "eval_biology_entropy": 1.2097046246528627,
359
+ "eval_biology_loss": 1.1880755424499512,
360
+ "eval_biology_mean_token_accuracy": 0.6989095420837402,
361
+ "eval_biology_num_tokens": 8506030.0,
362
+ "eval_biology_runtime": 19.7532,
363
+ "eval_biology_samples_per_second": 25.312,
364
+ "eval_biology_steps_per_second": 6.328,
365
+ "step": 300
366
+ },
367
+ {
368
+ "epoch": 0.48,
369
+ "eval_math_entropy": 0.7799157240390777,
370
+ "eval_math_loss": 0.7781939506530762,
371
+ "eval_math_mean_token_accuracy": 0.7959079008102417,
372
+ "eval_math_num_tokens": 8506030.0,
373
+ "eval_math_runtime": 24.5905,
374
+ "eval_math_samples_per_second": 20.333,
375
+ "eval_math_steps_per_second": 5.083,
376
+ "step": 300
377
+ },
378
+ {
379
+ "entropy": 0.7540466286242008,
380
+ "epoch": 0.496,
381
+ "grad_norm": 24.375,
382
+ "learning_rate": 6.18e-06,
383
+ "loss": 12.0453,
384
+ "mean_token_accuracy": 0.8007366862148046,
385
+ "num_tokens": 8788726.0,
386
+ "step": 310
387
+ },
388
+ {
389
+ "entropy": 0.7493760107085109,
390
+ "epoch": 0.512,
391
+ "grad_norm": 24.0,
392
+ "learning_rate": 6.380000000000001e-06,
393
+ "loss": 11.9382,
394
+ "mean_token_accuracy": 0.8025602623820305,
395
+ "num_tokens": 9078039.0,
396
+ "step": 320
397
+ },
398
+ {
399
+ "entropy": 0.7366351887583733,
400
+ "epoch": 0.528,
401
+ "grad_norm": 23.0,
402
+ "learning_rate": 6.5800000000000005e-06,
403
+ "loss": 11.7509,
404
+ "mean_token_accuracy": 0.8059428248554468,
405
+ "num_tokens": 9373860.0,
406
+ "step": 330
407
+ },
408
+ {
409
+ "entropy": 0.7388057533651591,
410
+ "epoch": 0.544,
411
+ "grad_norm": 24.125,
412
+ "learning_rate": 6.780000000000001e-06,
413
+ "loss": 11.827,
414
+ "mean_token_accuracy": 0.8040345013141632,
415
+ "num_tokens": 9660940.0,
416
+ "step": 340
417
+ },
418
+ {
419
+ "entropy": 0.7601569008082152,
420
+ "epoch": 0.56,
421
+ "grad_norm": 26.625,
422
+ "learning_rate": 6.98e-06,
423
+ "loss": 12.1324,
424
+ "mean_token_accuracy": 0.797855831682682,
425
+ "num_tokens": 9932302.0,
426
+ "step": 350
427
+ },
428
+ {
429
+ "entropy": 0.7458819771185518,
430
+ "epoch": 0.576,
431
+ "grad_norm": 22.375,
432
+ "learning_rate": 7.180000000000001e-06,
433
+ "loss": 11.9451,
434
+ "mean_token_accuracy": 0.8013740532100201,
435
+ "num_tokens": 10215462.0,
436
+ "step": 360
437
+ },
438
+ {
439
+ "entropy": 0.7404385067522525,
440
+ "epoch": 0.592,
441
+ "grad_norm": 21.625,
442
+ "learning_rate": 7.3800000000000005e-06,
443
+ "loss": 11.8309,
444
+ "mean_token_accuracy": 0.802900119498372,
445
+ "num_tokens": 10504396.0,
446
+ "step": 370
447
+ },
448
+ {
449
+ "entropy": 0.7532710742205382,
450
+ "epoch": 0.608,
451
+ "grad_norm": 27.75,
452
+ "learning_rate": 7.58e-06,
453
+ "loss": 12.0281,
454
+ "mean_token_accuracy": 0.801684994623065,
455
+ "num_tokens": 10793126.0,
456
+ "step": 380
457
+ },
458
+ {
459
+ "entropy": 0.7331796364858747,
460
+ "epoch": 0.624,
461
+ "grad_norm": 21.25,
462
+ "learning_rate": 7.78e-06,
463
+ "loss": 11.697,
464
+ "mean_token_accuracy": 0.8045222193002701,
465
+ "num_tokens": 11081768.0,
466
+ "step": 390
467
+ },
468
+ {
469
+ "entropy": 0.7177777705714107,
470
+ "epoch": 0.64,
471
+ "grad_norm": 20.0,
472
+ "learning_rate": 7.980000000000002e-06,
473
+ "loss": 11.4977,
474
+ "mean_token_accuracy": 0.8074667323380709,
475
+ "num_tokens": 11370320.0,
476
+ "step": 400
477
+ },
478
+ {
479
+ "epoch": 0.64,
480
+ "eval_biology_entropy": 1.199187099456787,
481
+ "eval_biology_loss": 1.194938063621521,
482
+ "eval_biology_mean_token_accuracy": 0.6980597639083862,
483
+ "eval_biology_num_tokens": 11370320.0,
484
+ "eval_biology_runtime": 19.7518,
485
+ "eval_biology_samples_per_second": 25.314,
486
+ "eval_biology_steps_per_second": 6.329,
487
+ "step": 400
488
+ },
489
+ {
490
+ "epoch": 0.64,
491
+ "eval_math_entropy": 0.7331400663852692,
492
+ "eval_math_loss": 0.746539831161499,
493
+ "eval_math_mean_token_accuracy": 0.8015342946052552,
494
+ "eval_math_num_tokens": 11370320.0,
495
+ "eval_math_runtime": 24.6063,
496
+ "eval_math_samples_per_second": 20.32,
497
+ "eval_math_steps_per_second": 5.08,
498
+ "step": 400
499
+ },
500
+ {
501
+ "entropy": 0.7210552679374814,
502
+ "epoch": 0.656,
503
+ "grad_norm": 20.75,
504
+ "learning_rate": 8.18e-06,
505
+ "loss": 11.4447,
506
+ "mean_token_accuracy": 0.8082952104508877,
507
+ "num_tokens": 11657835.0,
508
+ "step": 410
509
+ },
510
+ {
511
+ "entropy": 0.7324020706117154,
512
+ "epoch": 0.672,
513
+ "grad_norm": 22.5,
514
+ "learning_rate": 8.380000000000001e-06,
515
+ "loss": 11.7039,
516
+ "mean_token_accuracy": 0.8040592070668936,
517
+ "num_tokens": 11949262.0,
518
+ "step": 420
519
+ },
520
+ {
521
+ "entropy": 0.7221599837765098,
522
+ "epoch": 0.688,
523
+ "grad_norm": 19.5,
524
+ "learning_rate": 8.580000000000001e-06,
525
+ "loss": 11.5232,
526
+ "mean_token_accuracy": 0.8065517093986273,
527
+ "num_tokens": 12227640.0,
528
+ "step": 430
529
+ },
530
+ {
531
+ "entropy": 0.7129955545067788,
532
+ "epoch": 0.704,
533
+ "grad_norm": 19.0,
534
+ "learning_rate": 8.78e-06,
535
+ "loss": 11.4179,
536
+ "mean_token_accuracy": 0.8087377645075321,
537
+ "num_tokens": 12516641.0,
538
+ "step": 440
539
+ },
540
+ {
541
+ "entropy": 0.7289297079667449,
542
+ "epoch": 0.72,
543
+ "grad_norm": 23.125,
544
+ "learning_rate": 8.98e-06,
545
+ "loss": 11.5855,
546
+ "mean_token_accuracy": 0.8067171189934015,
547
+ "num_tokens": 12793343.0,
548
+ "step": 450
549
+ },
550
+ {
551
+ "entropy": 0.7121220523491502,
552
+ "epoch": 0.736,
553
+ "grad_norm": 21.0,
554
+ "learning_rate": 9.180000000000002e-06,
555
+ "loss": 11.3422,
556
+ "mean_token_accuracy": 0.8094062607735395,
557
+ "num_tokens": 13077981.0,
558
+ "step": 460
559
+ },
560
+ {
561
+ "entropy": 0.6984126020222903,
562
+ "epoch": 0.752,
563
+ "grad_norm": 18.625,
564
+ "learning_rate": 9.38e-06,
565
+ "loss": 11.0848,
566
+ "mean_token_accuracy": 0.811941733583808,
567
+ "num_tokens": 13358957.0,
568
+ "step": 470
569
+ },
570
+ {
571
+ "entropy": 0.677340486086905,
572
+ "epoch": 0.768,
573
+ "grad_norm": 20.875,
574
+ "learning_rate": 9.58e-06,
575
+ "loss": 10.8164,
576
+ "mean_token_accuracy": 0.8181491158902645,
577
+ "num_tokens": 13653412.0,
578
+ "step": 480
579
+ },
580
+ {
581
+ "entropy": 0.717779103666544,
582
+ "epoch": 0.784,
583
+ "grad_norm": 20.875,
584
+ "learning_rate": 9.780000000000001e-06,
585
+ "loss": 11.5008,
586
+ "mean_token_accuracy": 0.8082873310893774,
587
+ "num_tokens": 13940856.0,
588
+ "step": 490
589
+ },
590
+ {
591
+ "entropy": 0.7195664433762431,
592
+ "epoch": 0.8,
593
+ "grad_norm": 18.625,
594
+ "learning_rate": 9.980000000000001e-06,
595
+ "loss": 11.4516,
596
+ "mean_token_accuracy": 0.8075944270938635,
597
+ "num_tokens": 14230754.0,
598
+ "step": 500
599
+ },
600
+ {
601
+ "epoch": 0.8,
602
+ "eval_biology_entropy": 1.2116008014678956,
603
+ "eval_biology_loss": 1.1978343725204468,
604
+ "eval_biology_mean_token_accuracy": 0.6975936050415039,
605
+ "eval_biology_num_tokens": 14230754.0,
606
+ "eval_biology_runtime": 19.7814,
607
+ "eval_biology_samples_per_second": 25.276,
608
+ "eval_biology_steps_per_second": 6.319,
609
+ "step": 500
610
+ },
611
+ {
612
+ "epoch": 0.8,
613
+ "eval_math_entropy": 0.7416743865013122,
614
+ "eval_math_loss": 0.7205922603607178,
615
+ "eval_math_mean_token_accuracy": 0.8071010875701904,
616
+ "eval_math_num_tokens": 14230754.0,
617
+ "eval_math_runtime": 24.6035,
618
+ "eval_math_samples_per_second": 20.322,
619
+ "eval_math_steps_per_second": 5.081,
620
+ "step": 500
621
+ }
622
+ ],
623
+ "logging_steps": 10,
624
+ "max_steps": 10000,
625
+ "num_input_tokens_seen": 0,
626
+ "num_train_epochs": 16,
627
+ "save_steps": 500,
628
+ "stateful_callbacks": {
629
+ "TrainerControl": {
630
+ "args": {
631
+ "should_epoch_stop": false,
632
+ "should_evaluate": false,
633
+ "should_log": false,
634
+ "should_save": true,
635
+ "should_training_stop": false
636
+ },
637
+ "attributes": {}
638
+ }
639
+ },
640
+ "total_flos": 1.448451560178856e+18,
641
+ "train_batch_size": 4,
642
+ "trial_name": null,
643
+ "trial_params": null
644
+ }
checkpoint-500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f396f9399661130b840fe67ee6cf6fbebf3ce2dda00e2436a3e92b80430c2a6e
3
+ size 6417