Hugging Face's logo

roonbug
/
2sybub3b

Safetensors
Model card Files
xet
 Community
roonbug commited on
Commit
71b93a3
·
verified ·
1 Parent(s): 0635160

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 +754 -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 sha256:8bc75a29a730c9e743cad013feda3b0991a913fafe787c58a1c6e20afad97723
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:f84785ec61046319686d271520fe89ee0953245965d652481070bc81456953fb
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:ccad032f2edeeb6d0f54b9a1a9fbee800d41b406dd5beaa6604b519666a06086
3
+ size 896607944
checkpoint-500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fde986f97ecc270a34049538e5bcb495edbc4e7f5914a76b3cad083de287dc73
3
+ size 14645
checkpoint-500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ddc00081d4f892d131a0d2b4d4827a214275362bfa62dfb8413f489c1bfad457
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 +1,754 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 0.9036113740876317,
14
+ "epoch": 0.016,
15
+ "grad_norm": 128.0,
16
+ "learning_rate": 6.000000000000001e-07,
17
+ "loss": 34.907,
18
+ "mean_token_accuracy": 0.6652832090854645,
19
+ "num_tokens": 338352.0,
20
+ "step": 10
21
+ },
22
+ {
23
+ "entropy": 0.9087308555841446,
24
+ "epoch": 0.032,
25
+ "grad_norm": 119.5,
26
+ "learning_rate": 1.2666666666666669e-06,
27
+ "loss": 34.745,
28
+ "mean_token_accuracy": 0.6691140927374363,
29
+ "num_tokens": 671193.0,
30
+ "step": 20
31
+ },
32
+ {
33
+ "entropy": 0.9059555266052485,
34
+ "epoch": 0.048,
35
+ "grad_norm": 109.5,
36
+ "learning_rate": 1.9333333333333336e-06,
37
+ "loss": 34.1928,
38
+ "mean_token_accuracy": 0.6685802672058344,
39
+ "num_tokens": 1001008.0,
40
+ "step": 30
41
+ },
42
+ {
43
+ "entropy": 0.9450886320322752,
44
+ "epoch": 0.064,
45
+ "grad_norm": 102.0,
46
+ "learning_rate": 2.6e-06,
47
+ "loss": 34.8515,
48
+ "mean_token_accuracy": 0.6614982625469565,
49
+ "num_tokens": 1331208.0,
50
+ "step": 40
51
+ },
52
+ {
53
+ "entropy": 0.9593638770282269,
54
+ "epoch": 0.08,
55
+ "grad_norm": 94.5,
56
+ "learning_rate": 3.266666666666667e-06,
57
+ "loss": 33.1046,
58
+ "mean_token_accuracy": 0.6680185409262777,
59
+ "num_tokens": 1667532.0,
60
+ "step": 50
61
+ },
62
+ {
63
+ "entropy": 0.9606023650616408,
64
+ "epoch": 0.096,
65
+ "grad_norm": 80.0,
66
+ "learning_rate": 3.9333333333333335e-06,
67
+ "loss": 31.2876,
68
+ "mean_token_accuracy": 0.678196857497096,
69
+ "num_tokens": 2007176.0,
70
+ "step": 60
71
+ },
72
+ {
73
+ "entropy": 1.069316839799285,
74
+ "epoch": 0.112,
75
+ "grad_norm": 63.25,
76
+ "learning_rate": 4.600000000000001e-06,
77
+ "loss": 30.7512,
78
+ "mean_token_accuracy": 0.6743977405130863,
79
+ "num_tokens": 2342259.0,
80
+ "step": 70
81
+ },
82
+ {
83
+ "entropy": 1.2628379821777345,
84
+ "epoch": 0.128,
85
+ "grad_norm": 48.0,
86
+ "learning_rate": 5.2666666666666665e-06,
87
+ "loss": 30.3935,
88
+ "mean_token_accuracy": 0.6708028431981802,
89
+ "num_tokens": 2663987.0,
90
+ "step": 80
91
+ },
92
+ {
93
+ "entropy": 1.3055987607687711,
94
+ "epoch": 0.144,
95
+ "grad_norm": 32.25,
96
+ "learning_rate": 5.933333333333335e-06,
97
+ "loss": 27.3167,
98
+ "mean_token_accuracy": 0.6920903529971838,
99
+ "num_tokens": 2997824.0,
100
+ "step": 90
101
+ },
102
+ {
103
+ "entropy": 1.3579398691654205,
104
+ "epoch": 0.16,
105
+ "grad_norm": 24.625,
106
+ "learning_rate": 6.600000000000001e-06,
107
+ "loss": 25.5362,
108
+ "mean_token_accuracy": 0.7021430663764476,
109
+ "num_tokens": 3330597.0,
110
+ "step": 100
111
+ },
112
+ {
113
+ "epoch": 0.16,
114
+ "eval_biology_entropy": 3.549320913314819,
115
+ "eval_biology_loss": 4.242035388946533,
116
+ "eval_biology_mean_token_accuracy": 0.4143077299594879,
117
+ "eval_biology_num_tokens": 3330597.0,
118
+ "eval_biology_runtime": 55.9019,
119
+ "eval_biology_samples_per_second": 8.944,
120
+ "eval_biology_steps_per_second": 2.236,
121
+ "step": 100
122
+ },
123
+ {
124
+ "epoch": 0.16,
125
+ "eval_chemistry_entropy": 1.9644288024902343,
126
+ "eval_chemistry_loss": 2.1523141860961914,
127
+ "eval_chemistry_mean_token_accuracy": 0.635781112909317,
128
+ "eval_chemistry_num_tokens": 3330597.0,
129
+ "eval_chemistry_runtime": 69.6126,
130
+ "eval_chemistry_samples_per_second": 7.183,
131
+ "eval_chemistry_steps_per_second": 1.796,
132
+ "step": 100
133
+ },
134
+ {
135
+ "epoch": 0.16,
136
+ "eval_math_entropy": 1.1708643288612366,
137
+ "eval_math_loss": 1.6400102376937866,
138
+ "eval_math_mean_token_accuracy": 0.7028316712379455,
139
+ "eval_math_num_tokens": 3330597.0,
140
+ "eval_math_runtime": 71.6855,
141
+ "eval_math_samples_per_second": 6.975,
142
+ "eval_math_steps_per_second": 1.744,
143
+ "step": 100
144
+ },
145
+ {
146
+ "epoch": 0.16,
147
+ "eval_physics_entropy": 1.388730420589447,
148
+ "eval_physics_loss": 1.5701876878738403,
149
+ "eval_physics_mean_token_accuracy": 0.7025446329116821,
150
+ "eval_physics_num_tokens": 3330597.0,
151
+ "eval_physics_runtime": 83.1014,
152
+ "eval_physics_samples_per_second": 6.017,
153
+ "eval_physics_steps_per_second": 1.504,
154
+ "step": 100
155
+ },
156
+ {
157
+ "entropy": 1.3951868526637554,
158
+ "epoch": 0.176,
159
+ "grad_norm": 23.0,
160
+ "learning_rate": 7.266666666666668e-06,
161
+ "loss": 24.797,
162
+ "mean_token_accuracy": 0.7061628632247448,
163
+ "num_tokens": 3658264.0,
164
+ "step": 110
165
+ },
166
+ {
167
+ "entropy": 1.3387951213866471,
168
+ "epoch": 0.192,
169
+ "grad_norm": 15.375,
170
+ "learning_rate": 7.933333333333334e-06,
171
+ "loss": 23.7842,
172
+ "mean_token_accuracy": 0.713041927665472,
173
+ "num_tokens": 3995568.0,
174
+ "step": 120
175
+ },
176
+ {
177
+ "entropy": 1.3034498687833547,
178
+ "epoch": 0.208,
179
+ "grad_norm": 17.5,
180
+ "learning_rate": 8.6e-06,
181
+ "loss": 22.4839,
182
+ "mean_token_accuracy": 0.7252091735601425,
183
+ "num_tokens": 4321436.0,
184
+ "step": 130
185
+ },
186
+ {
187
+ "entropy": 1.2866257604211568,
188
+ "epoch": 0.224,
189
+ "grad_norm": 14.6875,
190
+ "learning_rate": 9.266666666666667e-06,
191
+ "loss": 22.0837,
192
+ "mean_token_accuracy": 0.7270880017429591,
193
+ "num_tokens": 4648491.0,
194
+ "step": 140
195
+ },
196
+ {
197
+ "entropy": 1.268964756652713,
198
+ "epoch": 0.24,
199
+ "grad_norm": 14.25,
200
+ "learning_rate": 9.933333333333334e-06,
201
+ "loss": 21.6129,
202
+ "mean_token_accuracy": 0.7301765695214272,
203
+ "num_tokens": 4986175.0,
204
+ "step": 150
205
+ },
206
+ {
207
+ "entropy": 1.2642716094851494,
208
+ "epoch": 0.256,
209
+ "grad_norm": 15.5625,
210
+ "learning_rate": 1.0600000000000002e-05,
211
+ "loss": 21.2618,
212
+ "mean_token_accuracy": 0.734415826946497,
213
+ "num_tokens": 5329320.0,
214
+ "step": 160
215
+ },
216
+ {
217
+ "entropy": 1.2717196103185415,
218
+ "epoch": 0.272,
219
+ "grad_norm": 17.5,
220
+ "learning_rate": 1.1266666666666668e-05,
221
+ "loss": 21.4982,
222
+ "mean_token_accuracy": 0.7297716859728098,
223
+ "num_tokens": 5658796.0,
224
+ "step": 170
225
+ },
226
+ {
227
+ "entropy": 1.2510493908077478,
228
+ "epoch": 0.288,
229
+ "grad_norm": 14.5625,
230
+ "learning_rate": 1.1933333333333335e-05,
231
+ "loss": 21.0954,
232
+ "mean_token_accuracy": 0.7358391582965851,
233
+ "num_tokens": 5980201.0,
234
+ "step": 180
235
+ },
236
+ {
237
+ "entropy": 1.2463122855871915,
238
+ "epoch": 0.304,
239
+ "grad_norm": 19.5,
240
+ "learning_rate": 1.2600000000000001e-05,
241
+ "loss": 20.8303,
242
+ "mean_token_accuracy": 0.7361553307622671,
243
+ "num_tokens": 6298858.0,
244
+ "step": 190
245
+ },
246
+ {
247
+ "entropy": 1.215079977735877,
248
+ "epoch": 0.32,
249
+ "grad_norm": 15.875,
250
+ "learning_rate": 1.3266666666666668e-05,
251
+ "loss": 20.4326,
252
+ "mean_token_accuracy": 0.7394854992628097,
253
+ "num_tokens": 6622798.0,
254
+ "step": 200
255
+ },
256
+ {
257
+ "epoch": 0.32,
258
+ "eval_biology_entropy": 3.1598935775756836,
259
+ "eval_biology_loss": 3.6490557193756104,
260
+ "eval_biology_mean_token_accuracy": 0.45875698828697203,
261
+ "eval_biology_num_tokens": 6622798.0,
262
+ "eval_biology_runtime": 56.3027,
263
+ "eval_biology_samples_per_second": 8.881,
264
+ "eval_biology_steps_per_second": 2.22,
265
+ "step": 200
266
+ },
267
+ {
268
+ "epoch": 0.32,
269
+ "eval_chemistry_entropy": 1.6644244351387023,
270
+ "eval_chemistry_loss": 1.768973469734192,
271
+ "eval_chemistry_mean_token_accuracy": 0.6738141541481018,
272
+ "eval_chemistry_num_tokens": 6622798.0,
273
+ "eval_chemistry_runtime": 69.6913,
274
+ "eval_chemistry_samples_per_second": 7.174,
275
+ "eval_chemistry_steps_per_second": 1.794,
276
+ "step": 200
277
+ },
278
+ {
279
+ "epoch": 0.32,
280
+ "eval_math_entropy": 1.124363118648529,
281
+ "eval_math_loss": 1.4134961366653442,
282
+ "eval_math_mean_token_accuracy": 0.7205373592376709,
283
+ "eval_math_num_tokens": 6622798.0,
284
+ "eval_math_runtime": 71.6804,
285
+ "eval_math_samples_per_second": 6.975,
286
+ "eval_math_steps_per_second": 1.744,
287
+ "step": 200
288
+ },
289
+ {
290
+ "epoch": 0.32,
291
+ "eval_physics_entropy": 1.2113610072135925,
292
+ "eval_physics_loss": 1.2715808153152466,
293
+ "eval_physics_mean_token_accuracy": 0.7393231744766235,
294
+ "eval_physics_num_tokens": 6622798.0,
295
+ "eval_physics_runtime": 83.0035,
296
+ "eval_physics_samples_per_second": 6.024,
297
+ "eval_physics_steps_per_second": 1.506,
298
+ "step": 200
299
+ },
300
+ {
301
+ "entropy": 1.2057966101914643,
302
+ "epoch": 0.336,
303
+ "grad_norm": 13.875,
304
+ "learning_rate": 1.3933333333333334e-05,
305
+ "loss": 20.1637,
306
+ "mean_token_accuracy": 0.740947175398469,
307
+ "num_tokens": 6953611.0,
308
+ "step": 210
309
+ },
310
+ {
311
+ "entropy": 1.1682135988026858,
312
+ "epoch": 0.352,
313
+ "grad_norm": 14.625,
314
+ "learning_rate": 1.46e-05,
315
+ "loss": 19.5323,
316
+ "mean_token_accuracy": 0.748641949146986,
317
+ "num_tokens": 7290378.0,
318
+ "step": 220
319
+ },
320
+ {
321
+ "entropy": 1.1533767595887183,
322
+ "epoch": 0.368,
323
+ "grad_norm": 14.75,
324
+ "learning_rate": 1.5266666666666667e-05,
325
+ "loss": 19.3393,
326
+ "mean_token_accuracy": 0.7496714886277914,
327
+ "num_tokens": 7621082.0,
328
+ "step": 230
329
+ },
330
+ {
331
+ "entropy": 1.2154120814055205,
332
+ "epoch": 0.384,
333
+ "grad_norm": 13.875,
334
+ "learning_rate": 1.5933333333333336e-05,
335
+ "loss": 20.1519,
336
+ "mean_token_accuracy": 0.7383838955312967,
337
+ "num_tokens": 7955570.0,
338
+ "step": 240
339
+ },
340
+ {
341
+ "entropy": 1.162219974398613,
342
+ "epoch": 0.4,
343
+ "grad_norm": 17.0,
344
+ "learning_rate": 1.66e-05,
345
+ "loss": 19.3544,
346
+ "mean_token_accuracy": 0.7475786007940769,
347
+ "num_tokens": 8291049.0,
348
+ "step": 250
349
+ },
350
+ {
351
+ "entropy": 1.1398306384682655,
352
+ "epoch": 0.416,
353
+ "grad_norm": 17.0,
354
+ "learning_rate": 1.726666666666667e-05,
355
+ "loss": 18.9335,
356
+ "mean_token_accuracy": 0.7522856764495373,
357
+ "num_tokens": 8627310.0,
358
+ "step": 260
359
+ },
360
+ {
361
+ "entropy": 1.160587527230382,
362
+ "epoch": 0.432,
363
+ "grad_norm": 14.5,
364
+ "learning_rate": 1.7933333333333333e-05,
365
+ "loss": 19.2248,
366
+ "mean_token_accuracy": 0.7491097755730152,
367
+ "num_tokens": 8958371.0,
368
+ "step": 270
369
+ },
370
+ {
371
+ "entropy": 1.1003841817378999,
372
+ "epoch": 0.448,
373
+ "grad_norm": 14.0,
374
+ "learning_rate": 1.86e-05,
375
+ "loss": 18.3143,
376
+ "mean_token_accuracy": 0.7592806451022625,
377
+ "num_tokens": 9294388.0,
378
+ "step": 280
379
+ },
380
+ {
381
+ "entropy": 1.1066032487899065,
382
+ "epoch": 0.464,
383
+ "grad_norm": 15.375,
384
+ "learning_rate": 1.926666666666667e-05,
385
+ "loss": 18.3092,
386
+ "mean_token_accuracy": 0.7574895836412907,
387
+ "num_tokens": 9620366.0,
388
+ "step": 290
389
+ },
390
+ {
391
+ "entropy": 1.1134729530662297,
392
+ "epoch": 0.48,
393
+ "grad_norm": 14.875,
394
+ "learning_rate": 1.9933333333333334e-05,
395
+ "loss": 18.5931,
396
+ "mean_token_accuracy": 0.7540634006261826,
397
+ "num_tokens": 9955431.0,
398
+ "step": 300
399
+ },
400
+ {
401
+ "epoch": 0.48,
402
+ "eval_biology_entropy": 3.058082123756409,
403
+ "eval_biology_loss": 3.4944381713867188,
404
+ "eval_biology_mean_token_accuracy": 0.4677197203636169,
405
+ "eval_biology_num_tokens": 9955431.0,
406
+ "eval_biology_runtime": 56.129,
407
+ "eval_biology_samples_per_second": 8.908,
408
+ "eval_biology_steps_per_second": 2.227,
409
+ "step": 300
410
+ },
411
+ {
412
+ "epoch": 0.48,
413
+ "eval_chemistry_entropy": 1.584096737384796,
414
+ "eval_chemistry_loss": 1.6560814380645752,
415
+ "eval_chemistry_mean_token_accuracy": 0.6847201838493348,
416
+ "eval_chemistry_num_tokens": 9955431.0,
417
+ "eval_chemistry_runtime": 69.5947,
418
+ "eval_chemistry_samples_per_second": 7.184,
419
+ "eval_chemistry_steps_per_second": 1.796,
420
+ "step": 300
421
+ },
422
+ {
423
+ "epoch": 0.48,
424
+ "eval_math_entropy": 1.1037719421386718,
425
+ "eval_math_loss": 1.3507343530654907,
426
+ "eval_math_mean_token_accuracy": 0.7274009571075439,
427
+ "eval_math_num_tokens": 9955431.0,
428
+ "eval_math_runtime": 71.8212,
429
+ "eval_math_samples_per_second": 6.962,
430
+ "eval_math_steps_per_second": 1.74,
431
+ "step": 300
432
+ },
433
+ {
434
+ "epoch": 0.48,
435
+ "eval_physics_entropy": 1.1398421788215638,
436
+ "eval_physics_loss": 1.1733136177062988,
437
+ "eval_physics_mean_token_accuracy": 0.7520353999137879,
438
+ "eval_physics_num_tokens": 9955431.0,
439
+ "eval_physics_runtime": 83.0575,
440
+ "eval_physics_samples_per_second": 6.02,
441
+ "eval_physics_steps_per_second": 1.505,
442
+ "step": 300
443
+ },
444
+ {
445
+ "entropy": 1.1404459465295076,
446
+ "epoch": 0.496,
447
+ "grad_norm": 14.25,
448
+ "learning_rate": 1.9933333333333334e-05,
449
+ "loss": 18.8326,
450
+ "mean_token_accuracy": 0.7513993989676238,
451
+ "num_tokens": 10289883.0,
452
+ "step": 310
453
+ },
454
+ {
455
+ "entropy": 1.114009590819478,
456
+ "epoch": 0.512,
457
+ "grad_norm": 15.4375,
458
+ "learning_rate": 1.985925925925926e-05,
459
+ "loss": 18.526,
460
+ "mean_token_accuracy": 0.7558605384081603,
461
+ "num_tokens": 10619468.0,
462
+ "step": 320
463
+ },
464
+ {
465
+ "entropy": 1.126872155070305,
466
+ "epoch": 0.528,
467
+ "grad_norm": 13.5625,
468
+ "learning_rate": 1.9785185185185187e-05,
469
+ "loss": 18.6944,
470
+ "mean_token_accuracy": 0.75301427654922,
471
+ "num_tokens": 10944904.0,
472
+ "step": 330
473
+ },
474
+ {
475
+ "entropy": 1.11179842017591,
476
+ "epoch": 0.544,
477
+ "grad_norm": 14.625,
478
+ "learning_rate": 1.971111111111111e-05,
479
+ "loss": 18.3887,
480
+ "mean_token_accuracy": 0.7559159055352211,
481
+ "num_tokens": 11266409.0,
482
+ "step": 340
483
+ },
484
+ {
485
+ "entropy": 1.0923905447125435,
486
+ "epoch": 0.56,
487
+ "grad_norm": 13.5,
488
+ "learning_rate": 1.963703703703704e-05,
489
+ "loss": 18.0431,
490
+ "mean_token_accuracy": 0.7601441144943237,
491
+ "num_tokens": 11605544.0,
492
+ "step": 350
493
+ },
494
+ {
495
+ "entropy": 1.075535924360156,
496
+ "epoch": 0.576,
497
+ "grad_norm": 15.375,
498
+ "learning_rate": 1.9562962962962964e-05,
499
+ "loss": 17.9134,
500
+ "mean_token_accuracy": 0.762933773174882,
501
+ "num_tokens": 11931715.0,
502
+ "step": 360
503
+ },
504
+ {
505
+ "entropy": 1.0857708267867565,
506
+ "epoch": 0.592,
507
+ "grad_norm": 16.5,
508
+ "learning_rate": 1.948888888888889e-05,
509
+ "loss": 17.9806,
510
+ "mean_token_accuracy": 0.7590868677943945,
511
+ "num_tokens": 12257243.0,
512
+ "step": 370
513
+ },
514
+ {
515
+ "entropy": 1.077372272312641,
516
+ "epoch": 0.608,
517
+ "grad_norm": 14.25,
518
+ "learning_rate": 1.9414814814814817e-05,
519
+ "loss": 17.8473,
520
+ "mean_token_accuracy": 0.7623418286442757,
521
+ "num_tokens": 12583154.0,
522
+ "step": 380
523
+ },
524
+ {
525
+ "entropy": 1.101408040151,
526
+ "epoch": 0.624,
527
+ "grad_norm": 13.4375,
528
+ "learning_rate": 1.9340740740740743e-05,
529
+ "loss": 18.1987,
530
+ "mean_token_accuracy": 0.7578636299818754,
531
+ "num_tokens": 12905392.0,
532
+ "step": 390
533
+ },
534
+ {
535
+ "entropy": 1.0936459716409446,
536
+ "epoch": 0.64,
537
+ "grad_norm": 15.0,
538
+ "learning_rate": 1.926666666666667e-05,
539
+ "loss": 18.2164,
540
+ "mean_token_accuracy": 0.7581360317766667,
541
+ "num_tokens": 13232198.0,
542
+ "step": 400
543
+ },
544
+ {
545
+ "epoch": 0.64,
546
+ "eval_biology_entropy": 2.985183084487915,
547
+ "eval_biology_loss": 3.3934555053710938,
548
+ "eval_biology_mean_token_accuracy": 0.47690209436416625,
549
+ "eval_biology_num_tokens": 13232198.0,
550
+ "eval_biology_runtime": 56.2992,
551
+ "eval_biology_samples_per_second": 8.881,
552
+ "eval_biology_steps_per_second": 2.22,
553
+ "step": 400
554
+ },
555
+ {
556
+ "epoch": 0.64,
557
+ "eval_chemistry_entropy": 1.5230218110084535,
558
+ "eval_chemistry_loss": 1.5974879264831543,
559
+ "eval_chemistry_mean_token_accuracy": 0.6924252939224244,
560
+ "eval_chemistry_num_tokens": 13232198.0,
561
+ "eval_chemistry_runtime": 69.6914,
562
+ "eval_chemistry_samples_per_second": 7.174,
563
+ "eval_chemistry_steps_per_second": 1.794,
564
+ "step": 400
565
+ },
566
+ {
567
+ "epoch": 0.64,
568
+ "eval_math_entropy": 1.0690260066986084,
569
+ "eval_math_loss": 1.3214365243911743,
570
+ "eval_math_mean_token_accuracy": 0.7313346996307373,
571
+ "eval_math_num_tokens": 13232198.0,
572
+ "eval_math_runtime": 71.8154,
573
+ "eval_math_samples_per_second": 6.962,
574
+ "eval_math_steps_per_second": 1.741,
575
+ "step": 400
576
+ },
577
+ {
578
+ "epoch": 0.64,
579
+ "eval_physics_entropy": 1.079647916316986,
580
+ "eval_physics_loss": 1.122850775718689,
581
+ "eval_physics_mean_token_accuracy": 0.7587474570274353,
582
+ "eval_physics_num_tokens": 13232198.0,
583
+ "eval_physics_runtime": 83.3676,
584
+ "eval_physics_samples_per_second": 5.998,
585
+ "eval_physics_steps_per_second": 1.499,
586
+ "step": 400
587
+ },
588
+ {
589
+ "entropy": 1.0861666969954968,
590
+ "epoch": 0.656,
591
+ "grad_norm": 14.1875,
592
+ "learning_rate": 1.9192592592592593e-05,
593
+ "loss": 17.9599,
594
+ "mean_token_accuracy": 0.7603602163493633,
595
+ "num_tokens": 13575902.0,
596
+ "step": 410
597
+ },
598
+ {
599
+ "entropy": 1.0595566183328629,
600
+ "epoch": 0.672,
601
+ "grad_norm": 15.5625,
602
+ "learning_rate": 1.911851851851852e-05,
603
+ "loss": 17.6133,
604
+ "mean_token_accuracy": 0.7644100178033113,
605
+ "num_tokens": 13895997.0,
606
+ "step": 420
607
+ },
608
+ {
609
+ "entropy": 1.0779153060168027,
610
+ "epoch": 0.688,
611
+ "grad_norm": 13.625,
612
+ "learning_rate": 1.9044444444444446e-05,
613
+ "loss": 17.708,
614
+ "mean_token_accuracy": 0.7641124926507473,
615
+ "num_tokens": 14234888.0,
616
+ "step": 430
617
+ },
618
+ {
619
+ "entropy": 1.0761012937873602,
620
+ "epoch": 0.704,
621
+ "grad_norm": 14.4375,
622
+ "learning_rate": 1.8970370370370372e-05,
623
+ "loss": 17.9412,
624
+ "mean_token_accuracy": 0.7593970015645027,
625
+ "num_tokens": 14567908.0,
626
+ "step": 440
627
+ },
628
+ {
629
+ "entropy": 1.0635488711297512,
630
+ "epoch": 0.72,
631
+ "grad_norm": 15.5625,
632
+ "learning_rate": 1.8896296296296295e-05,
633
+ "loss": 17.5409,
634
+ "mean_token_accuracy": 0.7655271884053946,
635
+ "num_tokens": 14882927.0,
636
+ "step": 450
637
+ },
638
+ {
639
+ "entropy": 1.0563195772469043,
640
+ "epoch": 0.736,
641
+ "grad_norm": 15.8125,
642
+ "learning_rate": 1.8822222222222225e-05,
643
+ "loss": 17.5759,
644
+ "mean_token_accuracy": 0.7641396790742874,
645
+ "num_tokens": 15217342.0,
646
+ "step": 460
647
+ },
648
+ {
649
+ "entropy": 1.037678936868906,
650
+ "epoch": 0.752,
651
+ "grad_norm": 13.125,
652
+ "learning_rate": 1.874814814814815e-05,
653
+ "loss": 17.1107,
654
+ "mean_token_accuracy": 0.7690521724522114,
655
+ "num_tokens": 15547710.0,
656
+ "step": 470
657
+ },
658
+ {
659
+ "entropy": 1.0838343646377324,
660
+ "epoch": 0.768,
661
+ "grad_norm": 16.75,
662
+ "learning_rate": 1.8674074074074075e-05,
663
+ "loss": 18.0446,
664
+ "mean_token_accuracy": 0.759132880717516,
665
+ "num_tokens": 15877177.0,
666
+ "step": 480
667
+ },
668
+ {
669
+ "entropy": 1.0296639803797007,
670
+ "epoch": 0.784,
671
+ "grad_norm": 12.5,
672
+ "learning_rate": 1.86e-05,
673
+ "loss": 17.1015,
674
+ "mean_token_accuracy": 0.7697366833686828,
675
+ "num_tokens": 16219640.0,
676
+ "step": 490
677
+ },
678
+ {
679
+ "entropy": 1.0563473679125308,
680
+ "epoch": 0.8,
681
+ "grad_norm": 15.5,
682
+ "learning_rate": 1.8525925925925928e-05,
683
+ "loss": 17.5054,
684
+ "mean_token_accuracy": 0.7657136749476194,
685
+ "num_tokens": 16548261.0,
686
+ "step": 500
687
+ },
688
+ {
689
+ "epoch": 0.8,
690
+ "eval_biology_entropy": 2.9737762069702147,
691
+ "eval_biology_loss": 3.3795254230499268,
692
+ "eval_biology_mean_token_accuracy": 0.48070894145965576,
693
+ "eval_biology_num_tokens": 16548261.0,
694
+ "eval_biology_runtime": 56.18,
695
+ "eval_biology_samples_per_second": 8.9,
696
+ "eval_biology_steps_per_second": 2.225,
697
+ "step": 500
698
+ },
699
+ {
700
+ "epoch": 0.8,
701
+ "eval_chemistry_entropy": 1.5115246725082399,
702
+ "eval_chemistry_loss": 1.573546051979065,
703
+ "eval_chemistry_mean_token_accuracy": 0.6963776540756226,
704
+ "eval_chemistry_num_tokens": 16548261.0,
705
+ "eval_chemistry_runtime": 69.0701,
706
+ "eval_chemistry_samples_per_second": 7.239,
707
+ "eval_chemistry_steps_per_second": 1.81,
708
+ "step": 500
709
+ },
710
+ {
711
+ "epoch": 0.8,
712
+ "eval_math_entropy": 1.0560337038040162,
713
+ "eval_math_loss": 1.306509017944336,
714
+ "eval_math_mean_token_accuracy": 0.7335917987823486,
715
+ "eval_math_num_tokens": 16548261.0,
716
+ "eval_math_runtime": 71.2835,
717
+ "eval_math_samples_per_second": 7.014,
718
+ "eval_math_steps_per_second": 1.754,
719
+ "step": 500
720
+ },
721
+ {
722
+ "epoch": 0.8,
723
+ "eval_physics_entropy": 1.0595522990226747,
724
+ "eval_physics_loss": 1.0949721336364746,
725
+ "eval_physics_mean_token_accuracy": 0.7634064397811889,
726
+ "eval_physics_num_tokens": 16548261.0,
727
+ "eval_physics_runtime": 83.0981,
728
+ "eval_physics_samples_per_second": 6.017,
729
+ "eval_physics_steps_per_second": 1.504,
730
+ "step": 500
731
+ }
732
+ ],
733
+ "logging_steps": 10,
734
+ "max_steps": 3000,
735
+ "num_input_tokens_seen": 0,
736
+ "num_train_epochs": 5,
737
+ "save_steps": 500,
738
+ "stateful_callbacks": {
739
+ "TrainerControl": {
740
+ "args": {
741
+ "should_epoch_stop": false,
742
+ "should_evaluate": false,
743
+ "should_log": false,
744
+ "should_save": true,
745
+ "should_training_stop": false
746
+ },
747
+ "attributes": {}
748
+ }
749
+ },
750
+ "total_flos": 1.652179086470234e+18,
751
+ "train_batch_size": 4,
752
+ "trial_name": null,
753
+ "trial_params": null
754
+ }
checkpoint-500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99fae199773974506471525072c20d6088b07c2300f704c09a94a2b4fb6d778d
3
+ size 6417