Hugging Face's logo

roonbug
/
4d8dcm2s

Safetensors
Model card Files
xet
 Community
roonbug commited on
Commit
76cba56
·
verified ·
1 Parent(s): 0a2f0b4

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:e603b3c01a4472038b23e2648e9cbfc58a0e5cb78bb8f308b5d30197c80dade5
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 sha256:1b3e4e2546ee70cf873e22f8de36b93437fd00631a242ad106ff41c5f7a3a6bc
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 sha256:1e09b79f7abc1c42d6d2618b5a4ac2bf7925d76888b3458030ab9d81b59f1b4d
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 sha256:5b9da25976931e768dc672638f220e0d21881583fe79c06d2c691b9bb66e5043
3
+ size 14573116445
checkpoint-500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef1ca0663cded63c4fdab8db6a019437ec66a59d3f24841741882983a86072d8
3
+ size 14645
checkpoint-500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9527d8453c68ef03b0e54ce001d2d4068a1f05836577d59b7d4e4a5a8aa7f0a8
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": 0.2,
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": 1.0975568080320954,
14
+ "epoch": 0.004,
15
+ "grad_norm": 1928.0,
16
+ "learning_rate": 4.5000000000000003e-07,
17
+ "loss": 51.036,
18
+ "mean_token_accuracy": 0.5464097319170833,
19
+ "num_tokens": 51831.0,
20
+ "step": 10
21
+ },
22
+ {
23
+ "entropy": 1.2103175908327102,
24
+ "epoch": 0.008,
25
+ "grad_norm": 1012.0,
26
+ "learning_rate": 9.500000000000001e-07,
27
+ "loss": 47.7535,
28
+ "mean_token_accuracy": 0.549801179766655,
29
+ "num_tokens": 103172.0,
30
+ "step": 20
31
+ },
32
+ {
33
+ "entropy": 1.4519254021346568,
34
+ "epoch": 0.012,
35
+ "grad_norm": 1024.0,
36
+ "learning_rate": 1.45e-06,
37
+ "loss": 45.5399,
38
+ "mean_token_accuracy": 0.5487965773791075,
39
+ "num_tokens": 147915.0,
40
+ "step": 30
41
+ },
42
+ {
43
+ "entropy": 1.515171455964446,
44
+ "epoch": 0.016,
45
+ "grad_norm": 440.0,
46
+ "learning_rate": 1.9500000000000004e-06,
47
+ "loss": 40.0298,
48
+ "mean_token_accuracy": 0.5707932172343135,
49
+ "num_tokens": 195524.0,
50
+ "step": 40
51
+ },
52
+ {
53
+ "entropy": 1.529548142105341,
54
+ "epoch": 0.02,
55
+ "grad_norm": 328.0,
56
+ "learning_rate": 2.4500000000000003e-06,
57
+ "loss": 33.3046,
58
+ "mean_token_accuracy": 0.5914752574637532,
59
+ "num_tokens": 243747.0,
60
+ "step": 50
61
+ },
62
+ {
63
+ "entropy": 1.5671692434698343,
64
+ "epoch": 0.024,
65
+ "grad_norm": 241.0,
66
+ "learning_rate": 2.95e-06,
67
+ "loss": 30.3037,
68
+ "mean_token_accuracy": 0.6058902809396386,
69
+ "num_tokens": 296999.0,
70
+ "step": 60
71
+ },
72
+ {
73
+ "entropy": 1.549248830974102,
74
+ "epoch": 0.028,
75
+ "grad_norm": 235.0,
76
+ "learning_rate": 3.45e-06,
77
+ "loss": 29.0295,
78
+ "mean_token_accuracy": 0.6220679895952343,
79
+ "num_tokens": 342114.0,
80
+ "step": 70
81
+ },
82
+ {
83
+ "entropy": 1.42005056142807,
84
+ "epoch": 0.032,
85
+ "grad_norm": 167.0,
86
+ "learning_rate": 3.95e-06,
87
+ "loss": 26.6089,
88
+ "mean_token_accuracy": 0.6380933992564678,
89
+ "num_tokens": 390903.0,
90
+ "step": 80
91
+ },
92
+ {
93
+ "entropy": 1.5127948723733424,
94
+ "epoch": 0.036,
95
+ "grad_norm": 167.0,
96
+ "learning_rate": 4.450000000000001e-06,
97
+ "loss": 24.8172,
98
+ "mean_token_accuracy": 0.6465719180181623,
99
+ "num_tokens": 440359.0,
100
+ "step": 90
101
+ },
102
+ {
103
+ "entropy": 1.5200586583465339,
104
+ "epoch": 0.04,
105
+ "grad_norm": 142.0,
106
+ "learning_rate": 4.95e-06,
107
+ "loss": 24.207,
108
+ "mean_token_accuracy": 0.6497391745448112,
109
+ "num_tokens": 488481.0,
110
+ "step": 100
111
+ },
112
+ {
113
+ "epoch": 0.04,
114
+ "eval_biology_entropy": 1.4843950763940812,
115
+ "eval_biology_loss": 1.4266694784164429,
116
+ "eval_biology_mean_token_accuracy": 0.6619762739539147,
117
+ "eval_biology_num_tokens": 488481.0,
118
+ "eval_biology_runtime": 36.5287,
119
+ "eval_biology_samples_per_second": 13.688,
120
+ "eval_biology_steps_per_second": 13.688,
121
+ "step": 100
122
+ },
123
+ {
124
+ "epoch": 0.04,
125
+ "eval_chemistry_entropy": 1.091414896786213,
126
+ "eval_chemistry_loss": 1.0560342073440552,
127
+ "eval_chemistry_mean_token_accuracy": 0.7318743930459023,
128
+ "eval_chemistry_num_tokens": 488481.0,
129
+ "eval_chemistry_runtime": 39.96,
130
+ "eval_chemistry_samples_per_second": 12.513,
131
+ "eval_chemistry_steps_per_second": 12.513,
132
+ "step": 100
133
+ },
134
+ {
135
+ "epoch": 0.04,
136
+ "eval_math_entropy": 0.9103146089911461,
137
+ "eval_math_loss": 1.0316376686096191,
138
+ "eval_math_mean_token_accuracy": 0.7397149602174758,
139
+ "eval_math_num_tokens": 488481.0,
140
+ "eval_math_runtime": 41.0822,
141
+ "eval_math_samples_per_second": 12.171,
142
+ "eval_math_steps_per_second": 12.171,
143
+ "step": 100
144
+ },
145
+ {
146
+ "epoch": 0.04,
147
+ "eval_physics_entropy": 0.9860039167404174,
148
+ "eval_physics_loss": 0.9796567559242249,
149
+ "eval_physics_mean_token_accuracy": 0.7483667035698891,
150
+ "eval_physics_num_tokens": 488481.0,
151
+ "eval_physics_runtime": 45.3249,
152
+ "eval_physics_samples_per_second": 11.031,
153
+ "eval_physics_steps_per_second": 11.031,
154
+ "step": 100
155
+ },
156
+ {
157
+ "entropy": 1.444656402617693,
158
+ "epoch": 0.044,
159
+ "grad_norm": 121.0,
160
+ "learning_rate": 5.450000000000001e-06,
161
+ "loss": 22.4665,
162
+ "mean_token_accuracy": 0.6580426702275872,
163
+ "num_tokens": 537471.0,
164
+ "step": 110
165
+ },
166
+ {
167
+ "entropy": 1.3339354760944844,
168
+ "epoch": 0.048,
169
+ "grad_norm": 129.0,
170
+ "learning_rate": 5.950000000000001e-06,
171
+ "loss": 21.4085,
172
+ "mean_token_accuracy": 0.6670540764927864,
173
+ "num_tokens": 589868.0,
174
+ "step": 120
175
+ },
176
+ {
177
+ "entropy": 1.3668562419712544,
178
+ "epoch": 0.052,
179
+ "grad_norm": 124.5,
180
+ "learning_rate": 6.450000000000001e-06,
181
+ "loss": 21.7438,
182
+ "mean_token_accuracy": 0.6594993283972144,
183
+ "num_tokens": 639512.0,
184
+ "step": 130
185
+ },
186
+ {
187
+ "entropy": 1.3065196067094802,
188
+ "epoch": 0.056,
189
+ "grad_norm": 118.5,
190
+ "learning_rate": 6.95e-06,
191
+ "loss": 20.9541,
192
+ "mean_token_accuracy": 0.6683343779295683,
193
+ "num_tokens": 690754.0,
194
+ "step": 140
195
+ },
196
+ {
197
+ "entropy": 1.3160824533551931,
198
+ "epoch": 0.06,
199
+ "grad_norm": 115.5,
200
+ "learning_rate": 7.450000000000001e-06,
201
+ "loss": 21.525,
202
+ "mean_token_accuracy": 0.664486956782639,
203
+ "num_tokens": 737075.0,
204
+ "step": 150
205
+ },
206
+ {
207
+ "entropy": 1.3052487019449472,
208
+ "epoch": 0.064,
209
+ "grad_norm": 113.0,
210
+ "learning_rate": 7.950000000000002e-06,
211
+ "loss": 20.5473,
212
+ "mean_token_accuracy": 0.672793828509748,
213
+ "num_tokens": 791190.0,
214
+ "step": 160
215
+ },
216
+ {
217
+ "entropy": 1.29992272593081,
218
+ "epoch": 0.068,
219
+ "grad_norm": 100.0,
220
+ "learning_rate": 8.45e-06,
221
+ "loss": 21.0557,
222
+ "mean_token_accuracy": 0.6683677529916168,
223
+ "num_tokens": 841344.0,
224
+ "step": 170
225
+ },
226
+ {
227
+ "entropy": 1.3232017546892165,
228
+ "epoch": 0.072,
229
+ "grad_norm": 114.5,
230
+ "learning_rate": 8.95e-06,
231
+ "loss": 21.0887,
232
+ "mean_token_accuracy": 0.6686861604452133,
233
+ "num_tokens": 890075.0,
234
+ "step": 180
235
+ },
236
+ {
237
+ "entropy": 1.2425255428999662,
238
+ "epoch": 0.076,
239
+ "grad_norm": 106.0,
240
+ "learning_rate": 9.450000000000001e-06,
241
+ "loss": 19.9137,
242
+ "mean_token_accuracy": 0.6810434710234403,
243
+ "num_tokens": 940599.0,
244
+ "step": 190
245
+ },
246
+ {
247
+ "entropy": 1.256486358679831,
248
+ "epoch": 0.08,
249
+ "grad_norm": 117.5,
250
+ "learning_rate": 9.950000000000001e-06,
251
+ "loss": 20.0961,
252
+ "mean_token_accuracy": 0.6819845270365477,
253
+ "num_tokens": 989860.0,
254
+ "step": 200
255
+ },
256
+ {
257
+ "epoch": 0.08,
258
+ "eval_biology_entropy": 1.241776801943779,
259
+ "eval_biology_loss": 1.2506648302078247,
260
+ "eval_biology_mean_token_accuracy": 0.6855908244848251,
261
+ "eval_biology_num_tokens": 989860.0,
262
+ "eval_biology_runtime": 36.0247,
263
+ "eval_biology_samples_per_second": 13.879,
264
+ "eval_biology_steps_per_second": 13.879,
265
+ "step": 200
266
+ },
267
+ {
268
+ "epoch": 0.08,
269
+ "eval_chemistry_entropy": 0.9190568276047707,
270
+ "eval_chemistry_loss": 0.9248532652854919,
271
+ "eval_chemistry_mean_token_accuracy": 0.7533808985352516,
272
+ "eval_chemistry_num_tokens": 989860.0,
273
+ "eval_chemistry_runtime": 40.487,
274
+ "eval_chemistry_samples_per_second": 12.35,
275
+ "eval_chemistry_steps_per_second": 12.35,
276
+ "step": 200
277
+ },
278
+ {
279
+ "epoch": 0.08,
280
+ "eval_math_entropy": 0.7673128904104233,
281
+ "eval_math_loss": 0.9293037056922913,
282
+ "eval_math_mean_token_accuracy": 0.759642232298851,
283
+ "eval_math_num_tokens": 989860.0,
284
+ "eval_math_runtime": 41.5439,
285
+ "eval_math_samples_per_second": 12.035,
286
+ "eval_math_steps_per_second": 12.035,
287
+ "step": 200
288
+ },
289
+ {
290
+ "epoch": 0.08,
291
+ "eval_physics_entropy": 0.8287813304066658,
292
+ "eval_physics_loss": 0.8567265868186951,
293
+ "eval_physics_mean_token_accuracy": 0.770364968419075,
294
+ "eval_physics_num_tokens": 989860.0,
295
+ "eval_physics_runtime": 46.6277,
296
+ "eval_physics_samples_per_second": 10.723,
297
+ "eval_physics_steps_per_second": 10.723,
298
+ "step": 200
299
+ },
300
+ {
301
+ "entropy": 1.2084858063608408,
302
+ "epoch": 0.084,
303
+ "grad_norm": 99.5,
304
+ "learning_rate": 1.045e-05,
305
+ "loss": 19.469,
306
+ "mean_token_accuracy": 0.6908657155930996,
307
+ "num_tokens": 1041199.0,
308
+ "step": 210
309
+ },
310
+ {
311
+ "entropy": 1.3143941588699817,
312
+ "epoch": 0.088,
313
+ "grad_norm": 139.0,
314
+ "learning_rate": 1.095e-05,
315
+ "loss": 21.209,
316
+ "mean_token_accuracy": 0.670545787923038,
317
+ "num_tokens": 1087099.0,
318
+ "step": 220
319
+ },
320
+ {
321
+ "entropy": 1.2587614998221397,
322
+ "epoch": 0.092,
323
+ "grad_norm": 108.0,
324
+ "learning_rate": 1.145e-05,
325
+ "loss": 20.1256,
326
+ "mean_token_accuracy": 0.6842886801809073,
327
+ "num_tokens": 1135090.0,
328
+ "step": 230
329
+ },
330
+ {
331
+ "entropy": 1.2565513662993908,
332
+ "epoch": 0.096,
333
+ "grad_norm": 106.0,
334
+ "learning_rate": 1.195e-05,
335
+ "loss": 20.4044,
336
+ "mean_token_accuracy": 0.6776034630835056,
337
+ "num_tokens": 1181777.0,
338
+ "step": 240
339
+ },
340
+ {
341
+ "entropy": 1.2938017453998327,
342
+ "epoch": 0.1,
343
+ "grad_norm": 112.0,
344
+ "learning_rate": 1.2450000000000003e-05,
345
+ "loss": 20.3714,
346
+ "mean_token_accuracy": 0.6780621778219939,
347
+ "num_tokens": 1229242.0,
348
+ "step": 250
349
+ },
350
+ {
351
+ "entropy": 1.2320711638778448,
352
+ "epoch": 0.104,
353
+ "grad_norm": 94.5,
354
+ "learning_rate": 1.295e-05,
355
+ "loss": 20.052,
356
+ "mean_token_accuracy": 0.6799502000212669,
357
+ "num_tokens": 1283881.0,
358
+ "step": 260
359
+ },
360
+ {
361
+ "entropy": 1.2415869060903788,
362
+ "epoch": 0.108,
363
+ "grad_norm": 107.0,
364
+ "learning_rate": 1.3450000000000002e-05,
365
+ "loss": 19.9151,
366
+ "mean_token_accuracy": 0.6853171430528164,
367
+ "num_tokens": 1332507.0,
368
+ "step": 270
369
+ },
370
+ {
371
+ "entropy": 1.242128773033619,
372
+ "epoch": 0.112,
373
+ "grad_norm": 92.5,
374
+ "learning_rate": 1.3950000000000002e-05,
375
+ "loss": 19.9203,
376
+ "mean_token_accuracy": 0.685390866547823,
377
+ "num_tokens": 1385513.0,
378
+ "step": 280
379
+ },
380
+ {
381
+ "entropy": 1.2662223283201457,
382
+ "epoch": 0.116,
383
+ "grad_norm": 100.5,
384
+ "learning_rate": 1.4450000000000002e-05,
385
+ "loss": 20.3929,
386
+ "mean_token_accuracy": 0.6766710925847292,
387
+ "num_tokens": 1432251.0,
388
+ "step": 290
389
+ },
390
+ {
391
+ "entropy": 1.198459953069687,
392
+ "epoch": 0.12,
393
+ "grad_norm": 101.0,
394
+ "learning_rate": 1.4950000000000003e-05,
395
+ "loss": 19.3584,
396
+ "mean_token_accuracy": 0.68702756986022,
397
+ "num_tokens": 1483667.0,
398
+ "step": 300
399
+ },
400
+ {
401
+ "epoch": 0.12,
402
+ "eval_biology_entropy": 1.2469175053834916,
403
+ "eval_biology_loss": 1.2332121133804321,
404
+ "eval_biology_mean_token_accuracy": 0.6868851281404496,
405
+ "eval_biology_num_tokens": 1483667.0,
406
+ "eval_biology_runtime": 36.8445,
407
+ "eval_biology_samples_per_second": 13.571,
408
+ "eval_biology_steps_per_second": 13.571,
409
+ "step": 300
410
+ },
411
+ {
412
+ "epoch": 0.12,
413
+ "eval_chemistry_entropy": 0.9295264416337014,
414
+ "eval_chemistry_loss": 0.9073873162269592,
415
+ "eval_chemistry_mean_token_accuracy": 0.7570922123789787,
416
+ "eval_chemistry_num_tokens": 1483667.0,
417
+ "eval_chemistry_runtime": 40.4179,
418
+ "eval_chemistry_samples_per_second": 12.371,
419
+ "eval_chemistry_steps_per_second": 12.371,
420
+ "step": 300
421
+ },
422
+ {
423
+ "epoch": 0.12,
424
+ "eval_math_entropy": 0.7820550892949104,
425
+ "eval_math_loss": 0.9024779200553894,
426
+ "eval_math_mean_token_accuracy": 0.7648742213249207,
427
+ "eval_math_num_tokens": 1483667.0,
428
+ "eval_math_runtime": 41.3487,
429
+ "eval_math_samples_per_second": 12.092,
430
+ "eval_math_steps_per_second": 12.092,
431
+ "step": 300
432
+ },
433
+ {
434
+ "epoch": 0.12,
435
+ "eval_physics_entropy": 0.8381168316006661,
436
+ "eval_physics_loss": 0.8357300758361816,
437
+ "eval_physics_mean_token_accuracy": 0.7741321437358856,
438
+ "eval_physics_num_tokens": 1483667.0,
439
+ "eval_physics_runtime": 46.0597,
440
+ "eval_physics_samples_per_second": 10.855,
441
+ "eval_physics_steps_per_second": 10.855,
442
+ "step": 300
443
+ },
444
+ {
445
+ "entropy": 1.268108993396163,
446
+ "epoch": 0.124,
447
+ "grad_norm": 87.0,
448
+ "learning_rate": 1.545e-05,
449
+ "loss": 20.169,
450
+ "mean_token_accuracy": 0.6794606490060687,
451
+ "num_tokens": 1534286.0,
452
+ "step": 310
453
+ },
454
+ {
455
+ "entropy": 1.2664785664528608,
456
+ "epoch": 0.128,
457
+ "grad_norm": 95.0,
458
+ "learning_rate": 1.595e-05,
459
+ "loss": 20.1543,
460
+ "mean_token_accuracy": 0.6814448466524482,
461
+ "num_tokens": 1582368.0,
462
+ "step": 320
463
+ },
464
+ {
465
+ "entropy": 1.2443583220243455,
466
+ "epoch": 0.132,
467
+ "grad_norm": 100.5,
468
+ "learning_rate": 1.645e-05,
469
+ "loss": 20.1453,
470
+ "mean_token_accuracy": 0.6805377397686243,
471
+ "num_tokens": 1630878.0,
472
+ "step": 330
473
+ },
474
+ {
475
+ "entropy": 1.295349196344614,
476
+ "epoch": 0.136,
477
+ "grad_norm": 84.5,
478
+ "learning_rate": 1.6950000000000002e-05,
479
+ "loss": 20.5474,
480
+ "mean_token_accuracy": 0.6783911313861608,
481
+ "num_tokens": 1677060.0,
482
+ "step": 340
483
+ },
484
+ {
485
+ "entropy": 1.2136471576988697,
486
+ "epoch": 0.14,
487
+ "grad_norm": 96.0,
488
+ "learning_rate": 1.7450000000000004e-05,
489
+ "loss": 19.7364,
490
+ "mean_token_accuracy": 0.6849099691957236,
491
+ "num_tokens": 1725186.0,
492
+ "step": 350
493
+ },
494
+ {
495
+ "entropy": 1.2611235719174148,
496
+ "epoch": 0.144,
497
+ "grad_norm": 96.5,
498
+ "learning_rate": 1.795e-05,
499
+ "loss": 20.2599,
500
+ "mean_token_accuracy": 0.6779890628531575,
501
+ "num_tokens": 1773764.0,
502
+ "step": 360
503
+ },
504
+ {
505
+ "entropy": 1.24961077272892,
506
+ "epoch": 0.148,
507
+ "grad_norm": 93.0,
508
+ "learning_rate": 1.845e-05,
509
+ "loss": 19.9901,
510
+ "mean_token_accuracy": 0.6829751830548048,
511
+ "num_tokens": 1823706.0,
512
+ "step": 370
513
+ },
514
+ {
515
+ "entropy": 1.2856210090219975,
516
+ "epoch": 0.152,
517
+ "grad_norm": 103.5,
518
+ "learning_rate": 1.8950000000000003e-05,
519
+ "loss": 20.5301,
520
+ "mean_token_accuracy": 0.6794548545032739,
521
+ "num_tokens": 1873167.0,
522
+ "step": 380
523
+ },
524
+ {
525
+ "entropy": 1.28184983022511,
526
+ "epoch": 0.156,
527
+ "grad_norm": 87.5,
528
+ "learning_rate": 1.9450000000000002e-05,
529
+ "loss": 20.4718,
530
+ "mean_token_accuracy": 0.6774068776518106,
531
+ "num_tokens": 1921484.0,
532
+ "step": 390
533
+ },
534
+ {
535
+ "entropy": 1.2587715853005648,
536
+ "epoch": 0.16,
537
+ "grad_norm": 90.5,
538
+ "learning_rate": 1.9950000000000004e-05,
539
+ "loss": 20.2765,
540
+ "mean_token_accuracy": 0.6795550292357803,
541
+ "num_tokens": 1970077.0,
542
+ "step": 400
543
+ },
544
+ {
545
+ "epoch": 0.16,
546
+ "eval_biology_entropy": 1.2585812546014785,
547
+ "eval_biology_loss": 1.2385821342468262,
548
+ "eval_biology_mean_token_accuracy": 0.6883087210059166,
549
+ "eval_biology_num_tokens": 1970077.0,
550
+ "eval_biology_runtime": 40.6484,
551
+ "eval_biology_samples_per_second": 12.301,
552
+ "eval_biology_steps_per_second": 12.301,
553
+ "step": 400
554
+ },
555
+ {
556
+ "epoch": 0.16,
557
+ "eval_chemistry_entropy": 0.9248176854848862,
558
+ "eval_chemistry_loss": 0.9106165766716003,
559
+ "eval_chemistry_mean_token_accuracy": 0.7573499709963799,
560
+ "eval_chemistry_num_tokens": 1970077.0,
561
+ "eval_chemistry_runtime": 40.4652,
562
+ "eval_chemistry_samples_per_second": 12.356,
563
+ "eval_chemistry_steps_per_second": 12.356,
564
+ "step": 400
565
+ },
566
+ {
567
+ "epoch": 0.16,
568
+ "eval_math_entropy": 0.7678850111961365,
569
+ "eval_math_loss": 0.897013783454895,
570
+ "eval_math_mean_token_accuracy": 0.7673858489990234,
571
+ "eval_math_num_tokens": 1970077.0,
572
+ "eval_math_runtime": 41.3756,
573
+ "eval_math_samples_per_second": 12.084,
574
+ "eval_math_steps_per_second": 12.084,
575
+ "step": 400
576
+ },
577
+ {
578
+ "epoch": 0.16,
579
+ "eval_physics_entropy": 0.8230015668869018,
580
+ "eval_physics_loss": 0.8317809104919434,
581
+ "eval_physics_mean_token_accuracy": 0.7749597446918488,
582
+ "eval_physics_num_tokens": 1970077.0,
583
+ "eval_physics_runtime": 45.5271,
584
+ "eval_physics_samples_per_second": 10.982,
585
+ "eval_physics_steps_per_second": 10.982,
586
+ "step": 400
587
+ },
588
+ {
589
+ "entropy": 1.2881209932267665,
590
+ "epoch": 0.164,
591
+ "grad_norm": 80.5,
592
+ "learning_rate": 1.9950000000000004e-05,
593
+ "loss": 20.7614,
594
+ "mean_token_accuracy": 0.6718229403719306,
595
+ "num_tokens": 2019053.0,
596
+ "step": 410
597
+ },
598
+ {
599
+ "entropy": 1.2165789466351271,
600
+ "epoch": 0.168,
601
+ "grad_norm": 75.5,
602
+ "learning_rate": 1.9894444444444446e-05,
603
+ "loss": 19.5986,
604
+ "mean_token_accuracy": 0.6869289472699165,
605
+ "num_tokens": 2066815.0,
606
+ "step": 420
607
+ },
608
+ {
609
+ "entropy": 1.2494619188830256,
610
+ "epoch": 0.172,
611
+ "grad_norm": 80.5,
612
+ "learning_rate": 1.983888888888889e-05,
613
+ "loss": 20.2759,
614
+ "mean_token_accuracy": 0.6808055322617292,
615
+ "num_tokens": 2116402.0,
616
+ "step": 430
617
+ },
618
+ {
619
+ "entropy": 1.214642644301057,
620
+ "epoch": 0.176,
621
+ "grad_norm": 81.5,
622
+ "learning_rate": 1.9783333333333334e-05,
623
+ "loss": 19.3769,
624
+ "mean_token_accuracy": 0.6881225043907762,
625
+ "num_tokens": 2168354.0,
626
+ "step": 440
627
+ },
628
+ {
629
+ "entropy": 1.2326878625899553,
630
+ "epoch": 0.18,
631
+ "grad_norm": 76.0,
632
+ "learning_rate": 1.972777777777778e-05,
633
+ "loss": 20.0507,
634
+ "mean_token_accuracy": 0.6823928710073233,
635
+ "num_tokens": 2219158.0,
636
+ "step": 450
637
+ },
638
+ {
639
+ "entropy": 1.2790402939543128,
640
+ "epoch": 0.184,
641
+ "grad_norm": 97.0,
642
+ "learning_rate": 1.9672222222222222e-05,
643
+ "loss": 20.4576,
644
+ "mean_token_accuracy": 0.6788654251024127,
645
+ "num_tokens": 2270617.0,
646
+ "step": 460
647
+ },
648
+ {
649
+ "entropy": 1.2233176242560149,
650
+ "epoch": 0.188,
651
+ "grad_norm": 76.0,
652
+ "learning_rate": 1.9616666666666668e-05,
653
+ "loss": 19.4471,
654
+ "mean_token_accuracy": 0.6852962577715516,
655
+ "num_tokens": 2322034.0,
656
+ "step": 470
657
+ },
658
+ {
659
+ "entropy": 1.284807311743498,
660
+ "epoch": 0.192,
661
+ "grad_norm": 97.0,
662
+ "learning_rate": 1.9561111111111114e-05,
663
+ "loss": 21.2325,
664
+ "mean_token_accuracy": 0.6681901682168245,
665
+ "num_tokens": 2365822.0,
666
+ "step": 480
667
+ },
668
+ {
669
+ "entropy": 1.250083539262414,
670
+ "epoch": 0.196,
671
+ "grad_norm": 93.5,
672
+ "learning_rate": 1.9505555555555556e-05,
673
+ "loss": 19.6203,
674
+ "mean_token_accuracy": 0.6853536464273929,
675
+ "num_tokens": 2413853.0,
676
+ "step": 490
677
+ },
678
+ {
679
+ "entropy": 1.2981292121112347,
680
+ "epoch": 0.2,
681
+ "grad_norm": 85.0,
682
+ "learning_rate": 1.9450000000000002e-05,
683
+ "loss": 20.8012,
684
+ "mean_token_accuracy": 0.6707905715331435,
685
+ "num_tokens": 2459876.0,
686
+ "step": 500
687
+ },
688
+ {
689
+ "epoch": 0.2,
690
+ "eval_biology_entropy": 1.2485336683392525,
691
+ "eval_biology_loss": 1.2281557321548462,
692
+ "eval_biology_mean_token_accuracy": 0.6897270123958588,
693
+ "eval_biology_num_tokens": 2459876.0,
694
+ "eval_biology_runtime": 38.8945,
695
+ "eval_biology_samples_per_second": 12.855,
696
+ "eval_biology_steps_per_second": 12.855,
697
+ "step": 500
698
+ },
699
+ {
700
+ "epoch": 0.2,
701
+ "eval_chemistry_entropy": 0.9259556359648704,
702
+ "eval_chemistry_loss": 0.9128974676132202,
703
+ "eval_chemistry_mean_token_accuracy": 0.7565669685602188,
704
+ "eval_chemistry_num_tokens": 2459876.0,
705
+ "eval_chemistry_runtime": 40.516,
706
+ "eval_chemistry_samples_per_second": 12.341,
707
+ "eval_chemistry_steps_per_second": 12.341,
708
+ "step": 500
709
+ },
710
+ {
711
+ "epoch": 0.2,
712
+ "eval_math_entropy": 0.7720121605396271,
713
+ "eval_math_loss": 0.9020798206329346,
714
+ "eval_math_mean_token_accuracy": 0.7660462435483932,
715
+ "eval_math_num_tokens": 2459876.0,
716
+ "eval_math_runtime": 62.8982,
717
+ "eval_math_samples_per_second": 7.949,
718
+ "eval_math_steps_per_second": 7.949,
719
+ "step": 500
720
+ },
721
+ {
722
+ "epoch": 0.2,
723
+ "eval_physics_entropy": 0.8241469451785087,
724
+ "eval_physics_loss": 0.832719624042511,
725
+ "eval_physics_mean_token_accuracy": 0.7746937532424927,
726
+ "eval_physics_num_tokens": 2459876.0,
727
+ "eval_physics_runtime": 76.8817,
728
+ "eval_physics_samples_per_second": 6.504,
729
+ "eval_physics_steps_per_second": 6.504,
730
+ "step": 500
731
+ }
732
+ ],
733
+ "logging_steps": 10,
734
+ "max_steps": 4000,
735
+ "num_input_tokens_seen": 0,
736
+ "num_train_epochs": 2,
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": 1.6494542652385037e+17,
751
+ "train_batch_size": 1,
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 sha256:205821b272c9e6374ffc464c4b2a6e8583b5f21d0bcc22110be9735c8da9d3f0
3
+ size 6481