Hugging Face's logo

roonbug
/
j5ej2dpo

Safetensors
Model card Files
xet
 Community
roonbug commited on
Commit
01042c4
·
verified ·
1 Parent(s): 9c58b6b

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 sha256:87eb2f7a36eaf795b0345a5db77849077e4fd41aea5b8b184b48544d3ea71b3d
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 sha256:2b9b5622b4bc3bb196c16cb69f467c964042e93dcdd265f294f3957e93483d3e
3
+ size 3187869259
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 sha256:63481e620bd287eefb582d20daf306c54c7268088fea6f19af8100945c7e3e7e
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": 0.22255605630668224,
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": 0.548704273160547,
14
+ "epoch": 0.004451121126133645,
15
+ "grad_norm": 1192.0,
16
+ "learning_rate": 1.8e-07,
17
+ "loss": 47.4036,
18
+ "mean_token_accuracy": 0.641148860566318,
19
+ "num_tokens": 63133.0,
20
+ "step": 10
21
+ },
22
+ {
23
+ "entropy": 0.534384406823665,
24
+ "epoch": 0.00890224225226729,
25
+ "grad_norm": 1328.0,
26
+ "learning_rate": 3.8e-07,
27
+ "loss": 50.4685,
28
+ "mean_token_accuracy": 0.6258877433836461,
29
+ "num_tokens": 128968.0,
30
+ "step": 20
31
+ },
32
+ {
33
+ "entropy": 0.537126874178648,
34
+ "epoch": 0.013353363378400934,
35
+ "grad_norm": 1320.0,
36
+ "learning_rate": 5.800000000000001e-07,
37
+ "loss": 49.4492,
38
+ "mean_token_accuracy": 0.6289473636075854,
39
+ "num_tokens": 193314.0,
40
+ "step": 30
41
+ },
42
+ {
43
+ "entropy": 0.5729353995062411,
44
+ "epoch": 0.01780448450453458,
45
+ "grad_norm": 1012.0,
46
+ "learning_rate": 7.8e-07,
47
+ "loss": 49.5744,
48
+ "mean_token_accuracy": 0.6247817112132907,
49
+ "num_tokens": 257189.0,
50
+ "step": 40
51
+ },
52
+ {
53
+ "entropy": 0.5564130786806345,
54
+ "epoch": 0.022255605630668224,
55
+ "grad_norm": 1096.0,
56
+ "learning_rate": 9.800000000000001e-07,
57
+ "loss": 44.4523,
58
+ "mean_token_accuracy": 0.6425775093957782,
59
+ "num_tokens": 324224.0,
60
+ "step": 50
61
+ },
62
+ {
63
+ "entropy": 0.590752430446446,
64
+ "epoch": 0.026706726756801868,
65
+ "grad_norm": 816.0,
66
+ "learning_rate": 1.1800000000000001e-06,
67
+ "loss": 43.7528,
68
+ "mean_token_accuracy": 0.6364219607785344,
69
+ "num_tokens": 394654.0,
70
+ "step": 60
71
+ },
72
+ {
73
+ "entropy": 0.6452584797516465,
74
+ "epoch": 0.031157847882935515,
75
+ "grad_norm": 600.0,
76
+ "learning_rate": 1.3800000000000001e-06,
77
+ "loss": 39.6529,
78
+ "mean_token_accuracy": 0.6429168622940779,
79
+ "num_tokens": 461658.0,
80
+ "step": 70
81
+ },
82
+ {
83
+ "entropy": 0.7016998372972012,
84
+ "epoch": 0.03560896900906916,
85
+ "grad_norm": 450.0,
86
+ "learning_rate": 1.5800000000000001e-06,
87
+ "loss": 35.508,
88
+ "mean_token_accuracy": 0.6540533849969506,
89
+ "num_tokens": 528285.0,
90
+ "step": 80
91
+ },
92
+ {
93
+ "entropy": 0.8241214359179139,
94
+ "epoch": 0.04006009013520281,
95
+ "grad_norm": 320.0,
96
+ "learning_rate": 1.7800000000000001e-06,
97
+ "loss": 34.6569,
98
+ "mean_token_accuracy": 0.6461574586108327,
99
+ "num_tokens": 595336.0,
100
+ "step": 90
101
+ },
102
+ {
103
+ "entropy": 1.0065209370106458,
104
+ "epoch": 0.04451121126133645,
105
+ "grad_norm": 208.0,
106
+ "learning_rate": 1.98e-06,
107
+ "loss": 31.2419,
108
+ "mean_token_accuracy": 0.6523716723546386,
109
+ "num_tokens": 663126.0,
110
+ "step": 100
111
+ },
112
+ {
113
+ "epoch": 0.04451121126133645,
114
+ "eval_biology_entropy": 1.5740653836727143,
115
+ "eval_biology_loss": 2.817919969558716,
116
+ "eval_biology_mean_token_accuracy": 0.5265991420149804,
117
+ "eval_biology_num_tokens": 663126.0,
118
+ "eval_biology_runtime": 49.0691,
119
+ "eval_biology_samples_per_second": 10.19,
120
+ "eval_biology_steps_per_second": 10.19,
121
+ "step": 100
122
+ },
123
+ {
124
+ "epoch": 0.04451121126133645,
125
+ "eval_chemistry_entropy": 1.1085504465103149,
126
+ "eval_chemistry_loss": 1.8567637205123901,
127
+ "eval_chemistry_mean_token_accuracy": 0.6607221029996871,
128
+ "eval_chemistry_num_tokens": 663126.0,
129
+ "eval_chemistry_runtime": 56.0546,
130
+ "eval_chemistry_samples_per_second": 8.92,
131
+ "eval_chemistry_steps_per_second": 8.92,
132
+ "step": 100
133
+ },
134
+ {
135
+ "epoch": 0.04451121126133645,
136
+ "eval_math_entropy": 0.7685319324731826,
137
+ "eval_math_loss": 1.6490864753723145,
138
+ "eval_math_mean_token_accuracy": 0.701621960401535,
139
+ "eval_math_num_tokens": 663126.0,
140
+ "eval_math_runtime": 57.7457,
141
+ "eval_math_samples_per_second": 8.659,
142
+ "eval_math_steps_per_second": 8.659,
143
+ "step": 100
144
+ },
145
+ {
146
+ "epoch": 0.04451121126133645,
147
+ "eval_physics_entropy": 0.9800559472441673,
148
+ "eval_physics_loss": 1.658416986465454,
149
+ "eval_physics_mean_token_accuracy": 0.6876975857019424,
150
+ "eval_physics_num_tokens": 663126.0,
151
+ "eval_physics_runtime": 64.5823,
152
+ "eval_physics_samples_per_second": 7.742,
153
+ "eval_physics_steps_per_second": 7.742,
154
+ "step": 100
155
+ },
156
+ {
157
+ "entropy": 1.2310039116069675,
158
+ "epoch": 0.048962332387470095,
159
+ "grad_norm": 129.0,
160
+ "learning_rate": 2.1800000000000003e-06,
161
+ "loss": 28.6754,
162
+ "mean_token_accuracy": 0.6632729232311249,
163
+ "num_tokens": 731060.0,
164
+ "step": 110
165
+ },
166
+ {
167
+ "entropy": 1.4431365672498941,
168
+ "epoch": 0.053413453513603736,
169
+ "grad_norm": 86.0,
170
+ "learning_rate": 2.38e-06,
171
+ "loss": 26.3421,
172
+ "mean_token_accuracy": 0.6755226453766227,
173
+ "num_tokens": 801635.0,
174
+ "step": 120
175
+ },
176
+ {
177
+ "entropy": 1.4806174896657467,
178
+ "epoch": 0.05786457463973738,
179
+ "grad_norm": 72.0,
180
+ "learning_rate": 2.5800000000000003e-06,
181
+ "loss": 25.6649,
182
+ "mean_token_accuracy": 0.6807831462472678,
183
+ "num_tokens": 867260.0,
184
+ "step": 130
185
+ },
186
+ {
187
+ "entropy": 1.511351894028485,
188
+ "epoch": 0.06231569576587103,
189
+ "grad_norm": 95.5,
190
+ "learning_rate": 2.7800000000000005e-06,
191
+ "loss": 25.1648,
192
+ "mean_token_accuracy": 0.6787236794829369,
193
+ "num_tokens": 931344.0,
194
+ "step": 140
195
+ },
196
+ {
197
+ "entropy": 1.4235187267884613,
198
+ "epoch": 0.06676681689200467,
199
+ "grad_norm": 57.25,
200
+ "learning_rate": 2.9800000000000003e-06,
201
+ "loss": 23.0409,
202
+ "mean_token_accuracy": 0.6943066673353314,
203
+ "num_tokens": 993586.0,
204
+ "step": 150
205
+ },
206
+ {
207
+ "entropy": 1.36321877297014,
208
+ "epoch": 0.07121793801813832,
209
+ "grad_norm": 52.75,
210
+ "learning_rate": 3.1800000000000005e-06,
211
+ "loss": 22.2695,
212
+ "mean_token_accuracy": 0.6991008354350925,
213
+ "num_tokens": 1060812.0,
214
+ "step": 160
215
+ },
216
+ {
217
+ "entropy": 1.2639451341703534,
218
+ "epoch": 0.07566905914427197,
219
+ "grad_norm": 55.0,
220
+ "learning_rate": 3.3800000000000007e-06,
221
+ "loss": 20.5103,
222
+ "mean_token_accuracy": 0.7160237554460764,
223
+ "num_tokens": 1131832.0,
224
+ "step": 170
225
+ },
226
+ {
227
+ "entropy": 1.3122306695207953,
228
+ "epoch": 0.08012018027040561,
229
+ "grad_norm": 43.25,
230
+ "learning_rate": 3.58e-06,
231
+ "loss": 20.6467,
232
+ "mean_token_accuracy": 0.7132720120251179,
233
+ "num_tokens": 1200650.0,
234
+ "step": 180
235
+ },
236
+ {
237
+ "entropy": 1.321925420127809,
238
+ "epoch": 0.08457130139653925,
239
+ "grad_norm": 52.25,
240
+ "learning_rate": 3.7800000000000002e-06,
241
+ "loss": 20.6542,
242
+ "mean_token_accuracy": 0.7144606521353125,
243
+ "num_tokens": 1265107.0,
244
+ "step": 190
245
+ },
246
+ {
247
+ "entropy": 1.304779276624322,
248
+ "epoch": 0.0890224225226729,
249
+ "grad_norm": 51.75,
250
+ "learning_rate": 3.980000000000001e-06,
251
+ "loss": 19.8821,
252
+ "mean_token_accuracy": 0.7192140439525246,
253
+ "num_tokens": 1327380.0,
254
+ "step": 200
255
+ },
256
+ {
257
+ "epoch": 0.0890224225226729,
258
+ "eval_biology_entropy": 1.7584778207540512,
259
+ "eval_biology_loss": 2.0029563903808594,
260
+ "eval_biology_mean_token_accuracy": 0.5894142254590988,
261
+ "eval_biology_num_tokens": 1327380.0,
262
+ "eval_biology_runtime": 48.4623,
263
+ "eval_biology_samples_per_second": 10.317,
264
+ "eval_biology_steps_per_second": 10.317,
265
+ "step": 200
266
+ },
267
+ {
268
+ "epoch": 0.0890224225226729,
269
+ "eval_chemistry_entropy": 1.2313024162650108,
270
+ "eval_chemistry_loss": 1.23660409450531,
271
+ "eval_chemistry_mean_token_accuracy": 0.7191956076025963,
272
+ "eval_chemistry_num_tokens": 1327380.0,
273
+ "eval_chemistry_runtime": 56.1215,
274
+ "eval_chemistry_samples_per_second": 8.909,
275
+ "eval_chemistry_steps_per_second": 8.909,
276
+ "step": 200
277
+ },
278
+ {
279
+ "epoch": 0.0890224225226729,
280
+ "eval_math_entropy": 0.9544051561951638,
281
+ "eval_math_loss": 1.1321452856063843,
282
+ "eval_math_mean_token_accuracy": 0.745829626917839,
283
+ "eval_math_num_tokens": 1327380.0,
284
+ "eval_math_runtime": 57.6893,
285
+ "eval_math_samples_per_second": 8.667,
286
+ "eval_math_steps_per_second": 8.667,
287
+ "step": 200
288
+ },
289
+ {
290
+ "epoch": 0.0890224225226729,
291
+ "eval_physics_entropy": 1.097377147436142,
292
+ "eval_physics_loss": 1.1237763166427612,
293
+ "eval_physics_mean_token_accuracy": 0.7436373601555825,
294
+ "eval_physics_num_tokens": 1327380.0,
295
+ "eval_physics_runtime": 64.6254,
296
+ "eval_physics_samples_per_second": 7.737,
297
+ "eval_physics_steps_per_second": 7.737,
298
+ "step": 200
299
+ },
300
+ {
301
+ "entropy": 1.2635900884866715,
302
+ "epoch": 0.09347354364880654,
303
+ "grad_norm": 54.5,
304
+ "learning_rate": 4.18e-06,
305
+ "loss": 20.5549,
306
+ "mean_token_accuracy": 0.7137404002249241,
307
+ "num_tokens": 1393379.0,
308
+ "step": 210
309
+ },
310
+ {
311
+ "entropy": 1.253603531420231,
312
+ "epoch": 0.09792466477494019,
313
+ "grad_norm": 45.25,
314
+ "learning_rate": 4.38e-06,
315
+ "loss": 19.3171,
316
+ "mean_token_accuracy": 0.7269632073119283,
317
+ "num_tokens": 1460130.0,
318
+ "step": 220
319
+ },
320
+ {
321
+ "entropy": 1.205259626172483,
322
+ "epoch": 0.10237578590107384,
323
+ "grad_norm": 50.5,
324
+ "learning_rate": 4.58e-06,
325
+ "loss": 18.3853,
326
+ "mean_token_accuracy": 0.7334358854219317,
327
+ "num_tokens": 1529183.0,
328
+ "step": 230
329
+ },
330
+ {
331
+ "entropy": 1.2351684581488371,
332
+ "epoch": 0.10682690702720747,
333
+ "grad_norm": 61.5,
334
+ "learning_rate": 4.78e-06,
335
+ "loss": 19.0263,
336
+ "mean_token_accuracy": 0.723250618390739,
337
+ "num_tokens": 1597405.0,
338
+ "step": 240
339
+ },
340
+ {
341
+ "entropy": 1.2351113758981227,
342
+ "epoch": 0.11127802815334112,
343
+ "grad_norm": 43.0,
344
+ "learning_rate": 4.980000000000001e-06,
345
+ "loss": 18.9974,
346
+ "mean_token_accuracy": 0.7199854323640466,
347
+ "num_tokens": 1661767.0,
348
+ "step": 250
349
+ },
350
+ {
351
+ "entropy": 1.1816746497526764,
352
+ "epoch": 0.11572914927947477,
353
+ "grad_norm": 45.75,
354
+ "learning_rate": 5.18e-06,
355
+ "loss": 18.6797,
356
+ "mean_token_accuracy": 0.7243335520848632,
357
+ "num_tokens": 1728207.0,
358
+ "step": 260
359
+ },
360
+ {
361
+ "entropy": 1.1266852594912051,
362
+ "epoch": 0.12018027040560841,
363
+ "grad_norm": 45.75,
364
+ "learning_rate": 5.380000000000001e-06,
365
+ "loss": 17.8377,
366
+ "mean_token_accuracy": 0.732505951821804,
367
+ "num_tokens": 1796055.0,
368
+ "step": 270
369
+ },
370
+ {
371
+ "entropy": 1.1155648654326797,
372
+ "epoch": 0.12463139153174206,
373
+ "grad_norm": 49.0,
374
+ "learning_rate": 5.580000000000001e-06,
375
+ "loss": 17.5397,
376
+ "mean_token_accuracy": 0.7362749936059118,
377
+ "num_tokens": 1859684.0,
378
+ "step": 280
379
+ },
380
+ {
381
+ "entropy": 1.1514084933325648,
382
+ "epoch": 0.1290825126578757,
383
+ "grad_norm": 51.0,
384
+ "learning_rate": 5.78e-06,
385
+ "loss": 18.5451,
386
+ "mean_token_accuracy": 0.723321282491088,
387
+ "num_tokens": 1921704.0,
388
+ "step": 290
389
+ },
390
+ {
391
+ "entropy": 1.1451087817549706,
392
+ "epoch": 0.13353363378400934,
393
+ "grad_norm": 42.5,
394
+ "learning_rate": 5.98e-06,
395
+ "loss": 18.2063,
396
+ "mean_token_accuracy": 0.7288043411448598,
397
+ "num_tokens": 1985766.0,
398
+ "step": 300
399
+ },
400
+ {
401
+ "epoch": 0.13353363378400934,
402
+ "eval_biology_entropy": 1.5927132781147957,
403
+ "eval_biology_loss": 1.7975378036499023,
404
+ "eval_biology_mean_token_accuracy": 0.6119548050761223,
405
+ "eval_biology_num_tokens": 1985766.0,
406
+ "eval_biology_runtime": 48.6056,
407
+ "eval_biology_samples_per_second": 10.287,
408
+ "eval_biology_steps_per_second": 10.287,
409
+ "step": 300
410
+ },
411
+ {
412
+ "epoch": 0.13353363378400934,
413
+ "eval_chemistry_entropy": 1.0956866154670715,
414
+ "eval_chemistry_loss": 1.0781587362289429,
415
+ "eval_chemistry_mean_token_accuracy": 0.7388799833655357,
416
+ "eval_chemistry_num_tokens": 1985766.0,
417
+ "eval_chemistry_runtime": 55.9983,
418
+ "eval_chemistry_samples_per_second": 8.929,
419
+ "eval_chemistry_steps_per_second": 8.929,
420
+ "step": 300
421
+ },
422
+ {
423
+ "epoch": 0.13353363378400934,
424
+ "eval_math_entropy": 0.8625128073692322,
425
+ "eval_math_loss": 1.0247634649276733,
426
+ "eval_math_mean_token_accuracy": 0.7569631568193436,
427
+ "eval_math_num_tokens": 1985766.0,
428
+ "eval_math_runtime": 57.7379,
429
+ "eval_math_samples_per_second": 8.66,
430
+ "eval_math_steps_per_second": 8.66,
431
+ "step": 300
432
+ },
433
+ {
434
+ "epoch": 0.13353363378400934,
435
+ "eval_physics_entropy": 0.9960197100043297,
436
+ "eval_physics_loss": 0.993407666683197,
437
+ "eval_physics_mean_token_accuracy": 0.7602200539112091,
438
+ "eval_physics_num_tokens": 1985766.0,
439
+ "eval_physics_runtime": 64.6649,
440
+ "eval_physics_samples_per_second": 7.732,
441
+ "eval_physics_steps_per_second": 7.732,
442
+ "step": 300
443
+ },
444
+ {
445
+ "entropy": 1.1051979778334498,
446
+ "epoch": 0.137984754910143,
447
+ "grad_norm": 49.5,
448
+ "learning_rate": 6.18e-06,
449
+ "loss": 17.5537,
450
+ "mean_token_accuracy": 0.7314090682193637,
451
+ "num_tokens": 2049041.0,
452
+ "step": 310
453
+ },
454
+ {
455
+ "entropy": 1.0414688095450402,
456
+ "epoch": 0.14243587603627664,
457
+ "grad_norm": 37.75,
458
+ "learning_rate": 6.380000000000001e-06,
459
+ "loss": 16.669,
460
+ "mean_token_accuracy": 0.7476102706044913,
461
+ "num_tokens": 2118068.0,
462
+ "step": 320
463
+ },
464
+ {
465
+ "entropy": 1.030502800643444,
466
+ "epoch": 0.14688699716241027,
467
+ "grad_norm": 46.5,
468
+ "learning_rate": 6.5800000000000005e-06,
469
+ "loss": 16.4295,
470
+ "mean_token_accuracy": 0.7489048192277551,
471
+ "num_tokens": 2186387.0,
472
+ "step": 330
473
+ },
474
+ {
475
+ "entropy": 1.0635908702388406,
476
+ "epoch": 0.15133811828854393,
477
+ "grad_norm": 42.0,
478
+ "learning_rate": 6.780000000000001e-06,
479
+ "loss": 17.235,
480
+ "mean_token_accuracy": 0.7380516286939383,
481
+ "num_tokens": 2252650.0,
482
+ "step": 340
483
+ },
484
+ {
485
+ "entropy": 1.0262556517496706,
486
+ "epoch": 0.15578923941467757,
487
+ "grad_norm": 41.0,
488
+ "learning_rate": 6.98e-06,
489
+ "loss": 16.1234,
490
+ "mean_token_accuracy": 0.7529099363833666,
491
+ "num_tokens": 2320358.0,
492
+ "step": 350
493
+ },
494
+ {
495
+ "entropy": 0.9883157461881638,
496
+ "epoch": 0.16024036054081123,
497
+ "grad_norm": 44.25,
498
+ "learning_rate": 7.180000000000001e-06,
499
+ "loss": 16.301,
500
+ "mean_token_accuracy": 0.7517775129526854,
501
+ "num_tokens": 2388824.0,
502
+ "step": 360
503
+ },
504
+ {
505
+ "entropy": 0.9906315876170992,
506
+ "epoch": 0.16469148166694486,
507
+ "grad_norm": 39.75,
508
+ "learning_rate": 7.3800000000000005e-06,
509
+ "loss": 15.5904,
510
+ "mean_token_accuracy": 0.760966083779931,
511
+ "num_tokens": 2456144.0,
512
+ "step": 370
513
+ },
514
+ {
515
+ "entropy": 1.0959791645407677,
516
+ "epoch": 0.1691426027930785,
517
+ "grad_norm": 37.5,
518
+ "learning_rate": 7.58e-06,
519
+ "loss": 17.7416,
520
+ "mean_token_accuracy": 0.7330215321853757,
521
+ "num_tokens": 2515325.0,
522
+ "step": 380
523
+ },
524
+ {
525
+ "entropy": 1.0152791179716587,
526
+ "epoch": 0.17359372391921216,
527
+ "grad_norm": 38.25,
528
+ "learning_rate": 7.78e-06,
529
+ "loss": 16.1438,
530
+ "mean_token_accuracy": 0.7489995956420898,
531
+ "num_tokens": 2580490.0,
532
+ "step": 390
533
+ },
534
+ {
535
+ "entropy": 1.015568952821195,
536
+ "epoch": 0.1780448450453458,
537
+ "grad_norm": 32.5,
538
+ "learning_rate": 7.980000000000002e-06,
539
+ "loss": 16.3986,
540
+ "mean_token_accuracy": 0.7477670447900892,
541
+ "num_tokens": 2644330.0,
542
+ "step": 400
543
+ },
544
+ {
545
+ "epoch": 0.1780448450453458,
546
+ "eval_biology_entropy": 1.4987669139504434,
547
+ "eval_biology_loss": 1.7285088300704956,
548
+ "eval_biology_mean_token_accuracy": 0.6236078375577927,
549
+ "eval_biology_num_tokens": 2644330.0,
550
+ "eval_biology_runtime": 48.385,
551
+ "eval_biology_samples_per_second": 10.334,
552
+ "eval_biology_steps_per_second": 10.334,
553
+ "step": 400
554
+ },
555
+ {
556
+ "epoch": 0.1780448450453458,
557
+ "eval_chemistry_entropy": 1.0116422069072724,
558
+ "eval_chemistry_loss": 1.0169917345046997,
559
+ "eval_chemistry_mean_token_accuracy": 0.7494727990031242,
560
+ "eval_chemistry_num_tokens": 2644330.0,
561
+ "eval_chemistry_runtime": 55.9816,
562
+ "eval_chemistry_samples_per_second": 8.932,
563
+ "eval_chemistry_steps_per_second": 8.932,
564
+ "step": 400
565
+ },
566
+ {
567
+ "epoch": 0.1780448450453458,
568
+ "eval_math_entropy": 0.830569740831852,
569
+ "eval_math_loss": 0.9943051934242249,
570
+ "eval_math_mean_token_accuracy": 0.7620642938613892,
571
+ "eval_math_num_tokens": 2644330.0,
572
+ "eval_math_runtime": 60.3723,
573
+ "eval_math_samples_per_second": 8.282,
574
+ "eval_math_steps_per_second": 8.282,
575
+ "step": 400
576
+ },
577
+ {
578
+ "epoch": 0.1780448450453458,
579
+ "eval_physics_entropy": 0.9355049659013748,
580
+ "eval_physics_loss": 0.94325852394104,
581
+ "eval_physics_mean_token_accuracy": 0.7684184198379517,
582
+ "eval_physics_num_tokens": 2644330.0,
583
+ "eval_physics_runtime": 65.391,
584
+ "eval_physics_samples_per_second": 7.646,
585
+ "eval_physics_steps_per_second": 7.646,
586
+ "step": 400
587
+ },
588
+ {
589
+ "entropy": 0.9359495824202895,
590
+ "epoch": 0.18249596617147945,
591
+ "grad_norm": 37.5,
592
+ "learning_rate": 8.18e-06,
593
+ "loss": 15.074,
594
+ "mean_token_accuracy": 0.7657283775508403,
595
+ "num_tokens": 2713234.0,
596
+ "step": 410
597
+ },
598
+ {
599
+ "entropy": 1.0310791170224547,
600
+ "epoch": 0.18694708729761308,
601
+ "grad_norm": 36.75,
602
+ "learning_rate": 8.380000000000001e-06,
603
+ "loss": 16.5812,
604
+ "mean_token_accuracy": 0.7489650597795844,
605
+ "num_tokens": 2783261.0,
606
+ "step": 420
607
+ },
608
+ {
609
+ "entropy": 1.0061013143509627,
610
+ "epoch": 0.19139820842374672,
611
+ "grad_norm": 34.25,
612
+ "learning_rate": 8.580000000000001e-06,
613
+ "loss": 16.2064,
614
+ "mean_token_accuracy": 0.7534871533513069,
615
+ "num_tokens": 2850170.0,
616
+ "step": 430
617
+ },
618
+ {
619
+ "entropy": 0.9983460938557982,
620
+ "epoch": 0.19584932954988038,
621
+ "grad_norm": 36.5,
622
+ "learning_rate": 8.78e-06,
623
+ "loss": 15.9251,
624
+ "mean_token_accuracy": 0.7519530924037099,
625
+ "num_tokens": 2913700.0,
626
+ "step": 440
627
+ },
628
+ {
629
+ "entropy": 0.9914884481579065,
630
+ "epoch": 0.200300450676014,
631
+ "grad_norm": 43.25,
632
+ "learning_rate": 8.98e-06,
633
+ "loss": 16.0176,
634
+ "mean_token_accuracy": 0.7508705228567123,
635
+ "num_tokens": 2981212.0,
636
+ "step": 450
637
+ },
638
+ {
639
+ "entropy": 0.9640501407906413,
640
+ "epoch": 0.20475157180214767,
641
+ "grad_norm": 43.75,
642
+ "learning_rate": 9.180000000000002e-06,
643
+ "loss": 15.5197,
644
+ "mean_token_accuracy": 0.7612664353102445,
645
+ "num_tokens": 3048973.0,
646
+ "step": 460
647
+ },
648
+ {
649
+ "entropy": 0.9837653011083602,
650
+ "epoch": 0.2092026929282813,
651
+ "grad_norm": 32.75,
652
+ "learning_rate": 9.38e-06,
653
+ "loss": 16.0462,
654
+ "mean_token_accuracy": 0.7515783123672009,
655
+ "num_tokens": 3117033.0,
656
+ "step": 470
657
+ },
658
+ {
659
+ "entropy": 0.9881603302434087,
660
+ "epoch": 0.21365381405441494,
661
+ "grad_norm": 35.5,
662
+ "learning_rate": 9.58e-06,
663
+ "loss": 15.6722,
664
+ "mean_token_accuracy": 0.7575565822422504,
665
+ "num_tokens": 3185255.0,
666
+ "step": 480
667
+ },
668
+ {
669
+ "entropy": 0.960085684619844,
670
+ "epoch": 0.2181049351805486,
671
+ "grad_norm": 43.5,
672
+ "learning_rate": 9.780000000000001e-06,
673
+ "loss": 15.732,
674
+ "mean_token_accuracy": 0.7576891854405403,
675
+ "num_tokens": 3253121.0,
676
+ "step": 490
677
+ },
678
+ {
679
+ "entropy": 0.9170863745734096,
680
+ "epoch": 0.22255605630668224,
681
+ "grad_norm": 39.5,
682
+ "learning_rate": 9.980000000000001e-06,
683
+ "loss": 14.4657,
684
+ "mean_token_accuracy": 0.7695914417505264,
685
+ "num_tokens": 3322823.0,
686
+ "step": 500
687
+ },
688
+ {
689
+ "epoch": 0.22255605630668224,
690
+ "eval_biology_entropy": 1.4002713676095009,
691
+ "eval_biology_loss": 1.6903640031814575,
692
+ "eval_biology_mean_token_accuracy": 0.6292068202495575,
693
+ "eval_biology_num_tokens": 3322823.0,
694
+ "eval_biology_runtime": 48.4234,
695
+ "eval_biology_samples_per_second": 10.326,
696
+ "eval_biology_steps_per_second": 10.326,
697
+ "step": 500
698
+ },
699
+ {
700
+ "epoch": 0.22255605630668224,
701
+ "eval_chemistry_entropy": 0.9267486627101899,
702
+ "eval_chemistry_loss": 0.9770777821540833,
703
+ "eval_chemistry_mean_token_accuracy": 0.7565804206728935,
704
+ "eval_chemistry_num_tokens": 3322823.0,
705
+ "eval_chemistry_runtime": 56.3359,
706
+ "eval_chemistry_samples_per_second": 8.875,
707
+ "eval_chemistry_steps_per_second": 8.875,
708
+ "step": 500
709
+ },
710
+ {
711
+ "epoch": 0.22255605630668224,
712
+ "eval_math_entropy": 0.7625217022895813,
713
+ "eval_math_loss": 0.9808707237243652,
714
+ "eval_math_mean_token_accuracy": 0.7647298012971878,
715
+ "eval_math_num_tokens": 3322823.0,
716
+ "eval_math_runtime": 57.8095,
717
+ "eval_math_samples_per_second": 8.649,
718
+ "eval_math_steps_per_second": 8.649,
719
+ "step": 500
720
+ },
721
+ {
722
+ "epoch": 0.22255605630668224,
723
+ "eval_physics_entropy": 0.8588201667070389,
724
+ "eval_physics_loss": 0.910319983959198,
725
+ "eval_physics_mean_token_accuracy": 0.7743916043043136,
726
+ "eval_physics_num_tokens": 3322823.0,
727
+ "eval_physics_runtime": 64.6422,
728
+ "eval_physics_samples_per_second": 7.735,
729
+ "eval_physics_steps_per_second": 7.735,
730
+ "step": 500
731
+ }
732
+ ],
733
+ "logging_steps": 10,
734
+ "max_steps": 10000,
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": 2.2280979081801686e+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 sha256:4b3edfed8df0c4397da38d1f8b970fbbf2479f63c7874fb7f01978dcc690947a
3
+ size 6481