Hugging Face's logo

YhangChen
/
iris-step100

Safetensors
multi_modality
Model card Files
xet
Community
YhangChen commited on
Commit
9b131b0
·
verified ·
1 Parent(s): 1e0e956

Initial model files upload

Browse files
Files changed (18) hide show
  1. config.json +85 -0
  2. latest +1 -0
  3. model-00001-of-00003.safetensors +3 -0
  4. model-00002-of-00003.safetensors +3 -0
  5. model-00003-of-00003.safetensors +3 -0
  6. model.safetensors.index.json +943 -0
  7. preprocessor_config.json +23 -0
  8. processor_config.json +12 -0
  9. rng_state_0.pth +3 -0
  10. rng_state_1.pth +3 -0
  11. rng_state_2.pth +3 -0
  12. rng_state_3.pth +3 -0
  13. special_tokens_map.json +34 -0
  14. tokenizer.json +0 -0
  15. tokenizer_config.json +0 -0
  16. trainer_state.json +1733 -0
  17. training_args.bin +3 -0
  18. zero_to_fp32.py +674 -0
config.json ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "deepseek-ai/Janus-Pro-7B",
3
+ "aligner_config": {
4
+ "cls": "MlpProjector",
5
+ "model_type": "aligner",
6
+ "params": {
7
+ "depth": 2,
8
+ "input_dim": 1024,
9
+ "n_embed": 4096,
10
+ "projector_type": "mlp_gelu"
11
+ }
12
+ },
13
+ "architectures": [
14
+ "MultiModalityCausalLM"
15
+ ],
16
+ "attention_bias": false,
17
+ "attention_dropout": 0.0,
18
+ "bf16": true,
19
+ "gen_aligner_config": {
20
+ "cls": "MlpProjector",
21
+ "model_type": "gen_aligner",
22
+ "params": {
23
+ "depth": 2,
24
+ "input_dim": 8,
25
+ "n_embed": 4096,
26
+ "projector_type": "mlp_gelu"
27
+ }
28
+ },
29
+ "gen_head_config": {
30
+ "cls": "vision_head",
31
+ "model_type": "gen_head",
32
+ "params": {
33
+ "image_token_embed": 4096,
34
+ "image_token_size": 16384,
35
+ "n_embed": 4096
36
+ }
37
+ },
38
+ "gen_vision_config": {
39
+ "cls": "VQ-16",
40
+ "model_type": "gen_vision",
41
+ "params": {
42
+ "image_token_size": 16384,
43
+ "n_embed": 8
44
+ }
45
+ },
46
+ "head_dim": 128,
47
+ "hidden_act": "silu",
48
+ "hidden_size": 4096,
49
+ "initializer_range": 0.02,
50
+ "intermediate_size": 11008,
51
+ "language_config": {
52
+ "_attn_implementation_autoset": true,
53
+ "bf16": true,
54
+ "max_position_embeddings": 16384,
55
+ "model_type": "llama",
56
+ "num_hidden_layers": 30,
57
+ "torch_dtype": "bfloat16",
58
+ "use_cache": false,
59
+ "vocab_size": 102400
60
+ },
61
+ "max_position_embeddings": 16384,
62
+ "mlp_bias": false,
63
+ "model_type": "multi_modality",
64
+ "num_attention_heads": 32,
65
+ "num_hidden_layers": 30,
66
+ "num_key_value_heads": 32,
67
+ "pretraining_tp": 1,
68
+ "rms_norm_eps": 1e-06,
69
+ "rope_scaling": null,
70
+ "rope_theta": 10000.0,
71
+ "torch_dtype": "bfloat16",
72
+ "transformers_version": "4.49.0",
73
+ "use_cache": true,
74
+ "vision_config": {
75
+ "cls": "CLIPVisionTower",
76
+ "model_type": "vision",
77
+ "params": {
78
+ "image_size": 384,
79
+ "model_name": "siglip_large_patch16_384",
80
+ "select_feature": "same",
81
+ "select_layer": -1
82
+ }
83
+ },
84
+ "vocab_size": 102400
85
+ }
latest ADDED
@@ -0,0 +1 @@
 
 
1
+ global_step100
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ef1d12de1f9b76de773e6ca21e139981b7f86ab966288a3a6d4a3588196c1cc
3
+ size 4916851534
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8fa7fd54b080d771855bde88451fbbebf46a9ebad587863174d4e81ef2f62c6
3
+ size 4947392496
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c1d37c50da069e73c7c0dd708dbd21441ceb66c95408b23ee5ee25dd6979886
3
+ size 4976742608
model.safetensors.index.json ADDED
@@ -0,0 +1,943 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 14840868118
4
+ },
5
+ "weight_map": {
6
+ "aligner.layers.0.bias": "model-00001-of-00003.safetensors",
7
+ "aligner.layers.0.weight": "model-00001-of-00003.safetensors",
8
+ "aligner.layers.2.bias": "model-00001-of-00003.safetensors",
9
+ "aligner.layers.2.weight": "model-00001-of-00003.safetensors",
10
+ "gen_aligner.layers.0.bias": "model-00001-of-00003.safetensors",
11
+ "gen_aligner.layers.0.weight": "model-00001-of-00003.safetensors",
12
+ "gen_aligner.layers.2.bias": "model-00001-of-00003.safetensors",
13
+ "gen_aligner.layers.2.weight": "model-00001-of-00003.safetensors",
14
+ "gen_embed.weight": "model-00001-of-00003.safetensors",
15
+ "gen_head.output_mlp_projector.bias": "model-00001-of-00003.safetensors",
16
+ "gen_head.output_mlp_projector.weight": "model-00001-of-00003.safetensors",
17
+ "gen_head.vision_head.bias": "model-00001-of-00003.safetensors",
18
+ "gen_head.vision_head.weight": "model-00001-of-00003.safetensors",
19
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.k.bias": "model-00001-of-00003.safetensors",
20
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.k.weight": "model-00001-of-00003.safetensors",
21
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.norm.bias": "model-00001-of-00003.safetensors",
22
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.norm.weight": "model-00001-of-00003.safetensors",
23
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.proj_out.bias": "model-00001-of-00003.safetensors",
24
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.proj_out.weight": "model-00001-of-00003.safetensors",
25
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.q.bias": "model-00001-of-00003.safetensors",
26
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.q.weight": "model-00001-of-00003.safetensors",
27
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.v.bias": "model-00001-of-00003.safetensors",
28
+ "gen_vision_model.decoder.conv_blocks.0.attn.0.v.weight": "model-00001-of-00003.safetensors",
29
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.k.bias": "model-00001-of-00003.safetensors",
30
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.k.weight": "model-00001-of-00003.safetensors",
31
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.norm.bias": "model-00001-of-00003.safetensors",
32
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.norm.weight": "model-00001-of-00003.safetensors",
33
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.proj_out.bias": "model-00001-of-00003.safetensors",
34
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.proj_out.weight": "model-00001-of-00003.safetensors",
35
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.q.bias": "model-00001-of-00003.safetensors",
36
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.q.weight": "model-00001-of-00003.safetensors",
37
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.v.bias": "model-00001-of-00003.safetensors",
38
+ "gen_vision_model.decoder.conv_blocks.0.attn.1.v.weight": "model-00001-of-00003.safetensors",
39
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.k.bias": "model-00001-of-00003.safetensors",
40
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.k.weight": "model-00001-of-00003.safetensors",
41
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.norm.bias": "model-00001-of-00003.safetensors",
42
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.norm.weight": "model-00001-of-00003.safetensors",
43
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.proj_out.bias": "model-00001-of-00003.safetensors",
44
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.proj_out.weight": "model-00001-of-00003.safetensors",
45
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.q.bias": "model-00001-of-00003.safetensors",
46
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.q.weight": "model-00001-of-00003.safetensors",
47
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.v.bias": "model-00001-of-00003.safetensors",
48
+ "gen_vision_model.decoder.conv_blocks.0.attn.2.v.weight": "model-00001-of-00003.safetensors",
49
+ "gen_vision_model.decoder.conv_blocks.0.res.0.conv1.bias": "model-00001-of-00003.safetensors",
50
+ "gen_vision_model.decoder.conv_blocks.0.res.0.conv1.weight": "model-00001-of-00003.safetensors",
51
+ "gen_vision_model.decoder.conv_blocks.0.res.0.conv2.bias": "model-00001-of-00003.safetensors",
52
+ "gen_vision_model.decoder.conv_blocks.0.res.0.conv2.weight": "model-00001-of-00003.safetensors",
53
+ "gen_vision_model.decoder.conv_blocks.0.res.0.norm1.bias": "model-00001-of-00003.safetensors",
54
+ "gen_vision_model.decoder.conv_blocks.0.res.0.norm1.weight": "model-00001-of-00003.safetensors",
55
+ "gen_vision_model.decoder.conv_blocks.0.res.0.norm2.bias": "model-00001-of-00003.safetensors",
56
+ "gen_vision_model.decoder.conv_blocks.0.res.0.norm2.weight": "model-00001-of-00003.safetensors",
57
+ "gen_vision_model.decoder.conv_blocks.0.res.1.conv1.bias": "model-00001-of-00003.safetensors",
58
+ "gen_vision_model.decoder.conv_blocks.0.res.1.conv1.weight": "model-00001-of-00003.safetensors",
59
+ "gen_vision_model.decoder.conv_blocks.0.res.1.conv2.bias": "model-00001-of-00003.safetensors",
60
+ "gen_vision_model.decoder.conv_blocks.0.res.1.conv2.weight": "model-00001-of-00003.safetensors",
61
+ "gen_vision_model.decoder.conv_blocks.0.res.1.norm1.bias": "model-00001-of-00003.safetensors",
62
+ "gen_vision_model.decoder.conv_blocks.0.res.1.norm1.weight": "model-00001-of-00003.safetensors",
63
+ "gen_vision_model.decoder.conv_blocks.0.res.1.norm2.bias": "model-00001-of-00003.safetensors",
64
+ "gen_vision_model.decoder.conv_blocks.0.res.1.norm2.weight": "model-00001-of-00003.safetensors",
65
+ "gen_vision_model.decoder.conv_blocks.0.res.2.conv1.bias": "model-00001-of-00003.safetensors",
66
+ "gen_vision_model.decoder.conv_blocks.0.res.2.conv1.weight": "model-00001-of-00003.safetensors",
67
+ "gen_vision_model.decoder.conv_blocks.0.res.2.conv2.bias": "model-00001-of-00003.safetensors",
68
+ "gen_vision_model.decoder.conv_blocks.0.res.2.conv2.weight": "model-00001-of-00003.safetensors",
69
+ "gen_vision_model.decoder.conv_blocks.0.res.2.norm1.bias": "model-00001-of-00003.safetensors",
70
+ "gen_vision_model.decoder.conv_blocks.0.res.2.norm1.weight": "model-00001-of-00003.safetensors",
71
+ "gen_vision_model.decoder.conv_blocks.0.res.2.norm2.bias": "model-00001-of-00003.safetensors",
72
+ "gen_vision_model.decoder.conv_blocks.0.res.2.norm2.weight": "model-00001-of-00003.safetensors",
73
+ "gen_vision_model.decoder.conv_blocks.0.upsample.conv.bias": "model-00001-of-00003.safetensors",
74
+ "gen_vision_model.decoder.conv_blocks.0.upsample.conv.weight": "model-00001-of-00003.safetensors",
75
+ "gen_vision_model.decoder.conv_blocks.1.res.0.conv1.bias": "model-00001-of-00003.safetensors",
76
+ "gen_vision_model.decoder.conv_blocks.1.res.0.conv1.weight": "model-00001-of-00003.safetensors",
77
+ "gen_vision_model.decoder.conv_blocks.1.res.0.conv2.bias": "model-00001-of-00003.safetensors",
78
+ "gen_vision_model.decoder.conv_blocks.1.res.0.conv2.weight": "model-00001-of-00003.safetensors",
79
+ "gen_vision_model.decoder.conv_blocks.1.res.0.nin_shortcut.bias": "model-00001-of-00003.safetensors",
80
+ "gen_vision_model.decoder.conv_blocks.1.res.0.nin_shortcut.weight": "model-00001-of-00003.safetensors",
81
+ "gen_vision_model.decoder.conv_blocks.1.res.0.norm1.bias": "model-00001-of-00003.safetensors",
82
+ "gen_vision_model.decoder.conv_blocks.1.res.0.norm1.weight": "model-00001-of-00003.safetensors",
83
+ "gen_vision_model.decoder.conv_blocks.1.res.0.norm2.bias": "model-00001-of-00003.safetensors",
84
+ "gen_vision_model.decoder.conv_blocks.1.res.0.norm2.weight": "model-00001-of-00003.safetensors",
85
+ "gen_vision_model.decoder.conv_blocks.1.res.1.conv1.bias": "model-00001-of-00003.safetensors",
86
+ "gen_vision_model.decoder.conv_blocks.1.res.1.conv1.weight": "model-00001-of-00003.safetensors",
87
+ "gen_vision_model.decoder.conv_blocks.1.res.1.conv2.bias": "model-00001-of-00003.safetensors",
88
+ "gen_vision_model.decoder.conv_blocks.1.res.1.conv2.weight": "model-00001-of-00003.safetensors",
89
+ "gen_vision_model.decoder.conv_blocks.1.res.1.norm1.bias": "model-00001-of-00003.safetensors",
90
+ "gen_vision_model.decoder.conv_blocks.1.res.1.norm1.weight": "model-00001-of-00003.safetensors",
91
+ "gen_vision_model.decoder.conv_blocks.1.res.1.norm2.bias": "model-00001-of-00003.safetensors",
92
+ "gen_vision_model.decoder.conv_blocks.1.res.1.norm2.weight": "model-00001-of-00003.safetensors",
93
+ "gen_vision_model.decoder.conv_blocks.1.res.2.conv1.bias": "model-00001-of-00003.safetensors",
94
+ "gen_vision_model.decoder.conv_blocks.1.res.2.conv1.weight": "model-00001-of-00003.safetensors",
95
+ "gen_vision_model.decoder.conv_blocks.1.res.2.conv2.bias": "model-00001-of-00003.safetensors",
96
+ "gen_vision_model.decoder.conv_blocks.1.res.2.conv2.weight": "model-00001-of-00003.safetensors",
97
+ "gen_vision_model.decoder.conv_blocks.1.res.2.norm1.bias": "model-00001-of-00003.safetensors",
98
+ "gen_vision_model.decoder.conv_blocks.1.res.2.norm1.weight": "model-00001-of-00003.safetensors",
99
+ "gen_vision_model.decoder.conv_blocks.1.res.2.norm2.bias": "model-00001-of-00003.safetensors",
100
+ "gen_vision_model.decoder.conv_blocks.1.res.2.norm2.weight": "model-00001-of-00003.safetensors",
101
+ "gen_vision_model.decoder.conv_blocks.1.upsample.conv.bias": "model-00001-of-00003.safetensors",
102
+ "gen_vision_model.decoder.conv_blocks.1.upsample.conv.weight": "model-00001-of-00003.safetensors",
103
+ "gen_vision_model.decoder.conv_blocks.2.res.0.conv1.bias": "model-00001-of-00003.safetensors",
104
+ "gen_vision_model.decoder.conv_blocks.2.res.0.conv1.weight": "model-00001-of-00003.safetensors",
105
+ "gen_vision_model.decoder.conv_blocks.2.res.0.conv2.bias": "model-00001-of-00003.safetensors",
106
+ "gen_vision_model.decoder.conv_blocks.2.res.0.conv2.weight": "model-00001-of-00003.safetensors",
107
+ "gen_vision_model.decoder.conv_blocks.2.res.0.norm1.bias": "model-00001-of-00003.safetensors",
108
+ "gen_vision_model.decoder.conv_blocks.2.res.0.norm1.weight": "model-00001-of-00003.safetensors",
109
+ "gen_vision_model.decoder.conv_blocks.2.res.0.norm2.bias": "model-00001-of-00003.safetensors",
110
+ "gen_vision_model.decoder.conv_blocks.2.res.0.norm2.weight": "model-00001-of-00003.safetensors",
111
+ "gen_vision_model.decoder.conv_blocks.2.res.1.conv1.bias": "model-00001-of-00003.safetensors",
112
+ "gen_vision_model.decoder.conv_blocks.2.res.1.conv1.weight": "model-00001-of-00003.safetensors",
113
+ "gen_vision_model.decoder.conv_blocks.2.res.1.conv2.bias": "model-00001-of-00003.safetensors",
114
+ "gen_vision_model.decoder.conv_blocks.2.res.1.conv2.weight": "model-00001-of-00003.safetensors",
115
+ "gen_vision_model.decoder.conv_blocks.2.res.1.norm1.bias": "model-00001-of-00003.safetensors",
116
+ "gen_vision_model.decoder.conv_blocks.2.res.1.norm1.weight": "model-00001-of-00003.safetensors",
117
+ "gen_vision_model.decoder.conv_blocks.2.res.1.norm2.bias": "model-00001-of-00003.safetensors",
118
+ "gen_vision_model.decoder.conv_blocks.2.res.1.norm2.weight": "model-00001-of-00003.safetensors",
119
+ "gen_vision_model.decoder.conv_blocks.2.res.2.conv1.bias": "model-00001-of-00003.safetensors",
120
+ "gen_vision_model.decoder.conv_blocks.2.res.2.conv1.weight": "model-00001-of-00003.safetensors",
121
+ "gen_vision_model.decoder.conv_blocks.2.res.2.conv2.bias": "model-00001-of-00003.safetensors",
122
+ "gen_vision_model.decoder.conv_blocks.2.res.2.conv2.weight": "model-00001-of-00003.safetensors",
123
+ "gen_vision_model.decoder.conv_blocks.2.res.2.norm1.bias": "model-00001-of-00003.safetensors",
124
+ "gen_vision_model.decoder.conv_blocks.2.res.2.norm1.weight": "model-00001-of-00003.safetensors",
125
+ "gen_vision_model.decoder.conv_blocks.2.res.2.norm2.bias": "model-00001-of-00003.safetensors",
126
+ "gen_vision_model.decoder.conv_blocks.2.res.2.norm2.weight": "model-00001-of-00003.safetensors",
127
+ "gen_vision_model.decoder.conv_blocks.2.upsample.conv.bias": "model-00001-of-00003.safetensors",
128
+ "gen_vision_model.decoder.conv_blocks.2.upsample.conv.weight": "model-00001-of-00003.safetensors",
129
+ "gen_vision_model.decoder.conv_blocks.3.res.0.conv1.bias": "model-00001-of-00003.safetensors",
130
+ "gen_vision_model.decoder.conv_blocks.3.res.0.conv1.weight": "model-00001-of-00003.safetensors",
131
+ "gen_vision_model.decoder.conv_blocks.3.res.0.conv2.bias": "model-00001-of-00003.safetensors",
132
+ "gen_vision_model.decoder.conv_blocks.3.res.0.conv2.weight": "model-00001-of-00003.safetensors",
133
+ "gen_vision_model.decoder.conv_blocks.3.res.0.nin_shortcut.bias": "model-00001-of-00003.safetensors",
134
+ "gen_vision_model.decoder.conv_blocks.3.res.0.nin_shortcut.weight": "model-00001-of-00003.safetensors",
135
+ "gen_vision_model.decoder.conv_blocks.3.res.0.norm1.bias": "model-00001-of-00003.safetensors",
136
+ "gen_vision_model.decoder.conv_blocks.3.res.0.norm1.weight": "model-00001-of-00003.safetensors",
137
+ "gen_vision_model.decoder.conv_blocks.3.res.0.norm2.bias": "model-00001-of-00003.safetensors",
138
+ "gen_vision_model.decoder.conv_blocks.3.res.0.norm2.weight": "model-00001-of-00003.safetensors",
139
+ "gen_vision_model.decoder.conv_blocks.3.res.1.conv1.bias": "model-00001-of-00003.safetensors",
140
+ "gen_vision_model.decoder.conv_blocks.3.res.1.conv1.weight": "model-00001-of-00003.safetensors",
141
+ "gen_vision_model.decoder.conv_blocks.3.res.1.conv2.bias": "model-00001-of-00003.safetensors",
142
+ "gen_vision_model.decoder.conv_blocks.3.res.1.conv2.weight": "model-00001-of-00003.safetensors",
143
+ "gen_vision_model.decoder.conv_blocks.3.res.1.norm1.bias": "model-00001-of-00003.safetensors",
144
+ "gen_vision_model.decoder.conv_blocks.3.res.1.norm1.weight": "model-00001-of-00003.safetensors",
145
+ "gen_vision_model.decoder.conv_blocks.3.res.1.norm2.bias": "model-00001-of-00003.safetensors",
146
+ "gen_vision_model.decoder.conv_blocks.3.res.1.norm2.weight": "model-00001-of-00003.safetensors",
147
+ "gen_vision_model.decoder.conv_blocks.3.res.2.conv1.bias": "model-00001-of-00003.safetensors",
148
+ "gen_vision_model.decoder.conv_blocks.3.res.2.conv1.weight": "model-00001-of-00003.safetensors",
149
+ "gen_vision_model.decoder.conv_blocks.3.res.2.conv2.bias": "model-00001-of-00003.safetensors",
150
+ "gen_vision_model.decoder.conv_blocks.3.res.2.conv2.weight": "model-00001-of-00003.safetensors",
151
+ "gen_vision_model.decoder.conv_blocks.3.res.2.norm1.bias": "model-00001-of-00003.safetensors",
152
+ "gen_vision_model.decoder.conv_blocks.3.res.2.norm1.weight": "model-00001-of-00003.safetensors",
153
+ "gen_vision_model.decoder.conv_blocks.3.res.2.norm2.bias": "model-00001-of-00003.safetensors",
154
+ "gen_vision_model.decoder.conv_blocks.3.res.2.norm2.weight": "model-00001-of-00003.safetensors",
155
+ "gen_vision_model.decoder.conv_blocks.3.upsample.conv.bias": "model-00001-of-00003.safetensors",
156
+ "gen_vision_model.decoder.conv_blocks.3.upsample.conv.weight": "model-00001-of-00003.safetensors",
157
+ "gen_vision_model.decoder.conv_blocks.4.res.0.conv1.bias": "model-00001-of-00003.safetensors",
158
+ "gen_vision_model.decoder.conv_blocks.4.res.0.conv1.weight": "model-00001-of-00003.safetensors",
159
+ "gen_vision_model.decoder.conv_blocks.4.res.0.conv2.bias": "model-00001-of-00003.safetensors",
160
+ "gen_vision_model.decoder.conv_blocks.4.res.0.conv2.weight": "model-00001-of-00003.safetensors",
161
+ "gen_vision_model.decoder.conv_blocks.4.res.0.norm1.bias": "model-00001-of-00003.safetensors",
162
+ "gen_vision_model.decoder.conv_blocks.4.res.0.norm1.weight": "model-00001-of-00003.safetensors",
163
+ "gen_vision_model.decoder.conv_blocks.4.res.0.norm2.bias": "model-00001-of-00003.safetensors",
164
+ "gen_vision_model.decoder.conv_blocks.4.res.0.norm2.weight": "model-00001-of-00003.safetensors",
165
+ "gen_vision_model.decoder.conv_blocks.4.res.1.conv1.bias": "model-00001-of-00003.safetensors",
166
+ "gen_vision_model.decoder.conv_blocks.4.res.1.conv1.weight": "model-00001-of-00003.safetensors",
167
+ "gen_vision_model.decoder.conv_blocks.4.res.1.conv2.bias": "model-00001-of-00003.safetensors",
168
+ "gen_vision_model.decoder.conv_blocks.4.res.1.conv2.weight": "model-00001-of-00003.safetensors",
169
+ "gen_vision_model.decoder.conv_blocks.4.res.1.norm1.bias": "model-00001-of-00003.safetensors",
170
+ "gen_vision_model.decoder.conv_blocks.4.res.1.norm1.weight": "model-00001-of-00003.safetensors",
171
+ "gen_vision_model.decoder.conv_blocks.4.res.1.norm2.bias": "model-00001-of-00003.safetensors",
172
+ "gen_vision_model.decoder.conv_blocks.4.res.1.norm2.weight": "model-00001-of-00003.safetensors",
173
+ "gen_vision_model.decoder.conv_blocks.4.res.2.conv1.bias": "model-00001-of-00003.safetensors",
174
+ "gen_vision_model.decoder.conv_blocks.4.res.2.conv1.weight": "model-00001-of-00003.safetensors",
175
+ "gen_vision_model.decoder.conv_blocks.4.res.2.conv2.bias": "model-00001-of-00003.safetensors",
176
+ "gen_vision_model.decoder.conv_blocks.4.res.2.conv2.weight": "model-00001-of-00003.safetensors",
177
+ "gen_vision_model.decoder.conv_blocks.4.res.2.norm1.bias": "model-00001-of-00003.safetensors",
178
+ "gen_vision_model.decoder.conv_blocks.4.res.2.norm1.weight": "model-00001-of-00003.safetensors",
179
+ "gen_vision_model.decoder.conv_blocks.4.res.2.norm2.bias": "model-00001-of-00003.safetensors",
180
+ "gen_vision_model.decoder.conv_blocks.4.res.2.norm2.weight": "model-00001-of-00003.safetensors",
181
+ "gen_vision_model.decoder.conv_in.bias": "model-00001-of-00003.safetensors",
182
+ "gen_vision_model.decoder.conv_in.weight": "model-00001-of-00003.safetensors",
183
+ "gen_vision_model.decoder.conv_out.bias": "model-00001-of-00003.safetensors",
184
+ "gen_vision_model.decoder.conv_out.weight": "model-00001-of-00003.safetensors",
185
+ "gen_vision_model.decoder.mid.0.conv1.bias": "model-00001-of-00003.safetensors",
186
+ "gen_vision_model.decoder.mid.0.conv1.weight": "model-00001-of-00003.safetensors",
187
+ "gen_vision_model.decoder.mid.0.conv2.bias": "model-00001-of-00003.safetensors",
188
+ "gen_vision_model.decoder.mid.0.conv2.weight": "model-00001-of-00003.safetensors",
189
+ "gen_vision_model.decoder.mid.0.norm1.bias": "model-00001-of-00003.safetensors",
190
+ "gen_vision_model.decoder.mid.0.norm1.weight": "model-00001-of-00003.safetensors",
191
+ "gen_vision_model.decoder.mid.0.norm2.bias": "model-00001-of-00003.safetensors",
192
+ "gen_vision_model.decoder.mid.0.norm2.weight": "model-00001-of-00003.safetensors",
193
+ "gen_vision_model.decoder.mid.1.k.bias": "model-00001-of-00003.safetensors",
194
+ "gen_vision_model.decoder.mid.1.k.weight": "model-00001-of-00003.safetensors",
195
+ "gen_vision_model.decoder.mid.1.norm.bias": "model-00001-of-00003.safetensors",
196
+ "gen_vision_model.decoder.mid.1.norm.weight": "model-00001-of-00003.safetensors",
197
+ "gen_vision_model.decoder.mid.1.proj_out.bias": "model-00001-of-00003.safetensors",
198
+ "gen_vision_model.decoder.mid.1.proj_out.weight": "model-00001-of-00003.safetensors",
199
+ "gen_vision_model.decoder.mid.1.q.bias": "model-00001-of-00003.safetensors",
200
+ "gen_vision_model.decoder.mid.1.q.weight": "model-00001-of-00003.safetensors",
201
+ "gen_vision_model.decoder.mid.1.v.bias": "model-00001-of-00003.safetensors",
202
+ "gen_vision_model.decoder.mid.1.v.weight": "model-00001-of-00003.safetensors",
203
+ "gen_vision_model.decoder.mid.2.conv1.bias": "model-00001-of-00003.safetensors",
204
+ "gen_vision_model.decoder.mid.2.conv1.weight": "model-00001-of-00003.safetensors",
205
+ "gen_vision_model.decoder.mid.2.conv2.bias": "model-00001-of-00003.safetensors",
206
+ "gen_vision_model.decoder.mid.2.conv2.weight": "model-00001-of-00003.safetensors",
207
+ "gen_vision_model.decoder.mid.2.norm1.bias": "model-00001-of-00003.safetensors",
208
+ "gen_vision_model.decoder.mid.2.norm1.weight": "model-00001-of-00003.safetensors",
209
+ "gen_vision_model.decoder.mid.2.norm2.bias": "model-00001-of-00003.safetensors",
210
+ "gen_vision_model.decoder.mid.2.norm2.weight": "model-00001-of-00003.safetensors",
211
+ "gen_vision_model.decoder.norm_out.bias": "model-00001-of-00003.safetensors",
212
+ "gen_vision_model.decoder.norm_out.weight": "model-00001-of-00003.safetensors",
213
+ "gen_vision_model.encoder.conv_blocks.0.downsample.conv.bias": "model-00001-of-00003.safetensors",
214
+ "gen_vision_model.encoder.conv_blocks.0.downsample.conv.weight": "model-00001-of-00003.safetensors",
215
+ "gen_vision_model.encoder.conv_blocks.0.res.0.conv1.bias": "model-00001-of-00003.safetensors",
216
+ "gen_vision_model.encoder.conv_blocks.0.res.0.conv1.weight": "model-00001-of-00003.safetensors",
217
+ "gen_vision_model.encoder.conv_blocks.0.res.0.conv2.bias": "model-00001-of-00003.safetensors",
218
+ "gen_vision_model.encoder.conv_blocks.0.res.0.conv2.weight": "model-00001-of-00003.safetensors",
219
+ "gen_vision_model.encoder.conv_blocks.0.res.0.norm1.bias": "model-00001-of-00003.safetensors",
220
+ "gen_vision_model.encoder.conv_blocks.0.res.0.norm1.weight": "model-00001-of-00003.safetensors",
221
+ "gen_vision_model.encoder.conv_blocks.0.res.0.norm2.bias": "model-00001-of-00003.safetensors",
222
+ "gen_vision_model.encoder.conv_blocks.0.res.0.norm2.weight": "model-00001-of-00003.safetensors",
223
+ "gen_vision_model.encoder.conv_blocks.0.res.1.conv1.bias": "model-00001-of-00003.safetensors",
224
+ "gen_vision_model.encoder.conv_blocks.0.res.1.conv1.weight": "model-00001-of-00003.safetensors",
225
+ "gen_vision_model.encoder.conv_blocks.0.res.1.conv2.bias": "model-00001-of-00003.safetensors",
226
+ "gen_vision_model.encoder.conv_blocks.0.res.1.conv2.weight": "model-00001-of-00003.safetensors",
227
+ "gen_vision_model.encoder.conv_blocks.0.res.1.norm1.bias": "model-00001-of-00003.safetensors",
228
+ "gen_vision_model.encoder.conv_blocks.0.res.1.norm1.weight": "model-00001-of-00003.safetensors",
229
+ "gen_vision_model.encoder.conv_blocks.0.res.1.norm2.bias": "model-00001-of-00003.safetensors",
230
+ "gen_vision_model.encoder.conv_blocks.0.res.1.norm2.weight": "model-00001-of-00003.safetensors",
231
+ "gen_vision_model.encoder.conv_blocks.1.downsample.conv.bias": "model-00001-of-00003.safetensors",
232
+ "gen_vision_model.encoder.conv_blocks.1.downsample.conv.weight": "model-00001-of-00003.safetensors",
233
+ "gen_vision_model.encoder.conv_blocks.1.res.0.conv1.bias": "model-00001-of-00003.safetensors",
234
+ "gen_vision_model.encoder.conv_blocks.1.res.0.conv1.weight": "model-00001-of-00003.safetensors",
235
+ "gen_vision_model.encoder.conv_blocks.1.res.0.conv2.bias": "model-00001-of-00003.safetensors",
236
+ "gen_vision_model.encoder.conv_blocks.1.res.0.conv2.weight": "model-00001-of-00003.safetensors",
237
+ "gen_vision_model.encoder.conv_blocks.1.res.0.norm1.bias": "model-00001-of-00003.safetensors",
238
+ "gen_vision_model.encoder.conv_blocks.1.res.0.norm1.weight": "model-00001-of-00003.safetensors",
239
+ "gen_vision_model.encoder.conv_blocks.1.res.0.norm2.bias": "model-00001-of-00003.safetensors",
240
+ "gen_vision_model.encoder.conv_blocks.1.res.0.norm2.weight": "model-00001-of-00003.safetensors",
241
+ "gen_vision_model.encoder.conv_blocks.1.res.1.conv1.bias": "model-00001-of-00003.safetensors",
242
+ "gen_vision_model.encoder.conv_blocks.1.res.1.conv1.weight": "model-00001-of-00003.safetensors",
243
+ "gen_vision_model.encoder.conv_blocks.1.res.1.conv2.bias": "model-00001-of-00003.safetensors",
244
+ "gen_vision_model.encoder.conv_blocks.1.res.1.conv2.weight": "model-00001-of-00003.safetensors",
245
+ "gen_vision_model.encoder.conv_blocks.1.res.1.norm1.bias": "model-00001-of-00003.safetensors",
246
+ "gen_vision_model.encoder.conv_blocks.1.res.1.norm1.weight": "model-00001-of-00003.safetensors",
247
+ "gen_vision_model.encoder.conv_blocks.1.res.1.norm2.bias": "model-00001-of-00003.safetensors",
248
+ "gen_vision_model.encoder.conv_blocks.1.res.1.norm2.weight": "model-00001-of-00003.safetensors",
249
+ "gen_vision_model.encoder.conv_blocks.2.downsample.conv.bias": "model-00001-of-00003.safetensors",
250
+ "gen_vision_model.encoder.conv_blocks.2.downsample.conv.weight": "model-00001-of-00003.safetensors",
251
+ "gen_vision_model.encoder.conv_blocks.2.res.0.conv1.bias": "model-00001-of-00003.safetensors",
252
+ "gen_vision_model.encoder.conv_blocks.2.res.0.conv1.weight": "model-00001-of-00003.safetensors",
253
+ "gen_vision_model.encoder.conv_blocks.2.res.0.conv2.bias": "model-00001-of-00003.safetensors",
254
+ "gen_vision_model.encoder.conv_blocks.2.res.0.conv2.weight": "model-00001-of-00003.safetensors",
255
+ "gen_vision_model.encoder.conv_blocks.2.res.0.nin_shortcut.bias": "model-00001-of-00003.safetensors",
256
+ "gen_vision_model.encoder.conv_blocks.2.res.0.nin_shortcut.weight": "model-00001-of-00003.safetensors",
257
+ "gen_vision_model.encoder.conv_blocks.2.res.0.norm1.bias": "model-00001-of-00003.safetensors",
258
+ "gen_vision_model.encoder.conv_blocks.2.res.0.norm1.weight": "model-00001-of-00003.safetensors",
259
+ "gen_vision_model.encoder.conv_blocks.2.res.0.norm2.bias": "model-00001-of-00003.safetensors",
260
+ "gen_vision_model.encoder.conv_blocks.2.res.0.norm2.weight": "model-00001-of-00003.safetensors",
261
+ "gen_vision_model.encoder.conv_blocks.2.res.1.conv1.bias": "model-00001-of-00003.safetensors",
262
+ "gen_vision_model.encoder.conv_blocks.2.res.1.conv1.weight": "model-00001-of-00003.safetensors",
263
+ "gen_vision_model.encoder.conv_blocks.2.res.1.conv2.bias": "model-00001-of-00003.safetensors",
264
+ "gen_vision_model.encoder.conv_blocks.2.res.1.conv2.weight": "model-00001-of-00003.safetensors",
265
+ "gen_vision_model.encoder.conv_blocks.2.res.1.norm1.bias": "model-00001-of-00003.safetensors",
266
+ "gen_vision_model.encoder.conv_blocks.2.res.1.norm1.weight": "model-00001-of-00003.safetensors",
267
+ "gen_vision_model.encoder.conv_blocks.2.res.1.norm2.bias": "model-00001-of-00003.safetensors",
268
+ "gen_vision_model.encoder.conv_blocks.2.res.1.norm2.weight": "model-00001-of-00003.safetensors",
269
+ "gen_vision_model.encoder.conv_blocks.3.downsample.conv.bias": "model-00001-of-00003.safetensors",
270
+ "gen_vision_model.encoder.conv_blocks.3.downsample.conv.weight": "model-00001-of-00003.safetensors",
271
+ "gen_vision_model.encoder.conv_blocks.3.res.0.conv1.bias": "model-00001-of-00003.safetensors",
272
+ "gen_vision_model.encoder.conv_blocks.3.res.0.conv1.weight": "model-00001-of-00003.safetensors",
273
+ "gen_vision_model.encoder.conv_blocks.3.res.0.conv2.bias": "model-00001-of-00003.safetensors",
274
+ "gen_vision_model.encoder.conv_blocks.3.res.0.conv2.weight": "model-00001-of-00003.safetensors",
275
+ "gen_vision_model.encoder.conv_blocks.3.res.0.norm1.bias": "model-00001-of-00003.safetensors",
276
+ "gen_vision_model.encoder.conv_blocks.3.res.0.norm1.weight": "model-00001-of-00003.safetensors",
277
+ "gen_vision_model.encoder.conv_blocks.3.res.0.norm2.bias": "model-00001-of-00003.safetensors",
278
+ "gen_vision_model.encoder.conv_blocks.3.res.0.norm2.weight": "model-00001-of-00003.safetensors",
279
+ "gen_vision_model.encoder.conv_blocks.3.res.1.conv1.bias": "model-00001-of-00003.safetensors",
280
+ "gen_vision_model.encoder.conv_blocks.3.res.1.conv1.weight": "model-00001-of-00003.safetensors",
281
+ "gen_vision_model.encoder.conv_blocks.3.res.1.conv2.bias": "model-00001-of-00003.safetensors",
282
+ "gen_vision_model.encoder.conv_blocks.3.res.1.conv2.weight": "model-00001-of-00003.safetensors",
283
+ "gen_vision_model.encoder.conv_blocks.3.res.1.norm1.bias": "model-00001-of-00003.safetensors",
284
+ "gen_vision_model.encoder.conv_blocks.3.res.1.norm1.weight": "model-00001-of-00003.safetensors",
285
+ "gen_vision_model.encoder.conv_blocks.3.res.1.norm2.bias": "model-00001-of-00003.safetensors",
286
+ "gen_vision_model.encoder.conv_blocks.3.res.1.norm2.weight": "model-00001-of-00003.safetensors",
287
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.k.bias": "model-00001-of-00003.safetensors",
288
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.k.weight": "model-00001-of-00003.safetensors",
289
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.norm.bias": "model-00001-of-00003.safetensors",
290
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.norm.weight": "model-00001-of-00003.safetensors",
291
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.proj_out.bias": "model-00001-of-00003.safetensors",
292
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.proj_out.weight": "model-00001-of-00003.safetensors",
293
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.q.bias": "model-00001-of-00003.safetensors",
294
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.q.weight": "model-00001-of-00003.safetensors",
295
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.v.bias": "model-00001-of-00003.safetensors",
296
+ "gen_vision_model.encoder.conv_blocks.4.attn.0.v.weight": "model-00001-of-00003.safetensors",
297
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.k.bias": "model-00001-of-00003.safetensors",
298
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.k.weight": "model-00001-of-00003.safetensors",
299
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.norm.bias": "model-00001-of-00003.safetensors",
300
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.norm.weight": "model-00001-of-00003.safetensors",
301
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.proj_out.bias": "model-00001-of-00003.safetensors",
302
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.proj_out.weight": "model-00001-of-00003.safetensors",
303
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.q.bias": "model-00001-of-00003.safetensors",
304
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.q.weight": "model-00001-of-00003.safetensors",
305
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.v.bias": "model-00001-of-00003.safetensors",
306
+ "gen_vision_model.encoder.conv_blocks.4.attn.1.v.weight": "model-00001-of-00003.safetensors",
307
+ "gen_vision_model.encoder.conv_blocks.4.res.0.conv1.bias": "model-00001-of-00003.safetensors",
308
+ "gen_vision_model.encoder.conv_blocks.4.res.0.conv1.weight": "model-00001-of-00003.safetensors",
309
+ "gen_vision_model.encoder.conv_blocks.4.res.0.conv2.bias": "model-00001-of-00003.safetensors",
310
+ "gen_vision_model.encoder.conv_blocks.4.res.0.conv2.weight": "model-00001-of-00003.safetensors",
311
+ "gen_vision_model.encoder.conv_blocks.4.res.0.nin_shortcut.bias": "model-00001-of-00003.safetensors",
312
+ "gen_vision_model.encoder.conv_blocks.4.res.0.nin_shortcut.weight": "model-00001-of-00003.safetensors",
313
+ "gen_vision_model.encoder.conv_blocks.4.res.0.norm1.bias": "model-00001-of-00003.safetensors",
314
+ "gen_vision_model.encoder.conv_blocks.4.res.0.norm1.weight": "model-00001-of-00003.safetensors",
315
+ "gen_vision_model.encoder.conv_blocks.4.res.0.norm2.bias": "model-00001-of-00003.safetensors",
316
+ "gen_vision_model.encoder.conv_blocks.4.res.0.norm2.weight": "model-00001-of-00003.safetensors",
317
+ "gen_vision_model.encoder.conv_blocks.4.res.1.conv1.bias": "model-00001-of-00003.safetensors",
318
+ "gen_vision_model.encoder.conv_blocks.4.res.1.conv1.weight": "model-00001-of-00003.safetensors",
319
+ "gen_vision_model.encoder.conv_blocks.4.res.1.conv2.bias": "model-00001-of-00003.safetensors",
320
+ "gen_vision_model.encoder.conv_blocks.4.res.1.conv2.weight": "model-00001-of-00003.safetensors",
321
+ "gen_vision_model.encoder.conv_blocks.4.res.1.norm1.bias": "model-00001-of-00003.safetensors",
322
+ "gen_vision_model.encoder.conv_blocks.4.res.1.norm1.weight": "model-00001-of-00003.safetensors",
323
+ "gen_vision_model.encoder.conv_blocks.4.res.1.norm2.bias": "model-00001-of-00003.safetensors",
324
+ "gen_vision_model.encoder.conv_blocks.4.res.1.norm2.weight": "model-00001-of-00003.safetensors",
325
+ "gen_vision_model.encoder.conv_in.bias": "model-00001-of-00003.safetensors",
326
+ "gen_vision_model.encoder.conv_in.weight": "model-00001-of-00003.safetensors",
327
+ "gen_vision_model.encoder.conv_out.bias": "model-00001-of-00003.safetensors",
328
+ "gen_vision_model.encoder.conv_out.weight": "model-00001-of-00003.safetensors",
329
+ "gen_vision_model.encoder.mid.0.conv1.bias": "model-00001-of-00003.safetensors",
330
+ "gen_vision_model.encoder.mid.0.conv1.weight": "model-00001-of-00003.safetensors",
331
+ "gen_vision_model.encoder.mid.0.conv2.bias": "model-00001-of-00003.safetensors",
332
+ "gen_vision_model.encoder.mid.0.conv2.weight": "model-00001-of-00003.safetensors",
333
+ "gen_vision_model.encoder.mid.0.norm1.bias": "model-00001-of-00003.safetensors",
334
+ "gen_vision_model.encoder.mid.0.norm1.weight": "model-00001-of-00003.safetensors",
335
+ "gen_vision_model.encoder.mid.0.norm2.bias": "model-00001-of-00003.safetensors",
336
+ "gen_vision_model.encoder.mid.0.norm2.weight": "model-00001-of-00003.safetensors",
337
+ "gen_vision_model.encoder.mid.1.k.bias": "model-00001-of-00003.safetensors",
338
+ "gen_vision_model.encoder.mid.1.k.weight": "model-00001-of-00003.safetensors",
339
+ "gen_vision_model.encoder.mid.1.norm.bias": "model-00001-of-00003.safetensors",
340
+ "gen_vision_model.encoder.mid.1.norm.weight": "model-00001-of-00003.safetensors",
341
+ "gen_vision_model.encoder.mid.1.proj_out.bias": "model-00001-of-00003.safetensors",
342
+ "gen_vision_model.encoder.mid.1.proj_out.weight": "model-00001-of-00003.safetensors",
343
+ "gen_vision_model.encoder.mid.1.q.bias": "model-00001-of-00003.safetensors",
344
+ "gen_vision_model.encoder.mid.1.q.weight": "model-00001-of-00003.safetensors",
345
+ "gen_vision_model.encoder.mid.1.v.bias": "model-00001-of-00003.safetensors",
346
+ "gen_vision_model.encoder.mid.1.v.weight": "model-00001-of-00003.safetensors",
347
+ "gen_vision_model.encoder.mid.2.conv1.bias": "model-00001-of-00003.safetensors",
348
+ "gen_vision_model.encoder.mid.2.conv1.weight": "model-00001-of-00003.safetensors",
349
+ "gen_vision_model.encoder.mid.2.conv2.bias": "model-00001-of-00003.safetensors",
350
+ "gen_vision_model.encoder.mid.2.conv2.weight": "model-00001-of-00003.safetensors",
351
+ "gen_vision_model.encoder.mid.2.norm1.bias": "model-00001-of-00003.safetensors",
352
+ "gen_vision_model.encoder.mid.2.norm1.weight": "model-00001-of-00003.safetensors",
353
+ "gen_vision_model.encoder.mid.2.norm2.bias": "model-00001-of-00003.safetensors",
354
+ "gen_vision_model.encoder.mid.2.norm2.weight": "model-00001-of-00003.safetensors",
355
+ "gen_vision_model.encoder.norm_out.bias": "model-00001-of-00003.safetensors",
356
+ "gen_vision_model.encoder.norm_out.weight": "model-00001-of-00003.safetensors",
357
+ "gen_vision_model.post_quant_conv.bias": "model-00001-of-00003.safetensors",
358
+ "gen_vision_model.post_quant_conv.weight": "model-00001-of-00003.safetensors",
359
+ "gen_vision_model.quant_conv.bias": "model-00001-of-00003.safetensors",
360
+ "gen_vision_model.quant_conv.weight": "model-00001-of-00003.safetensors",
361
+ "gen_vision_model.quantize.codebook_used": "model-00001-of-00003.safetensors",
362
+ "gen_vision_model.quantize.embedding.weight": "model-00001-of-00003.safetensors",
363
+ "language_model.lm_head.weight": "model-00003-of-00003.safetensors",
364
+ "language_model.model.embed_tokens.weight": "model-00001-of-00003.safetensors",
365
+ "language_model.model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
366
+ "language_model.model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
367
+ "language_model.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
368
+ "language_model.model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
369
+ "language_model.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
370
+ "language_model.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
371
+ "language_model.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
372
+ "language_model.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
373
+ "language_model.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
374
+ "language_model.model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
375
+ "language_model.model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
376
+ "language_model.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
377
+ "language_model.model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
378
+ "language_model.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
379
+ "language_model.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
380
+ "language_model.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
381
+ "language_model.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
382
+ "language_model.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
383
+ "language_model.model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
384
+ "language_model.model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
385
+ "language_model.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
386
+ "language_model.model.layers.10.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
387
+ "language_model.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
388
+ "language_model.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
389
+ "language_model.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
390
+ "language_model.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
391
+ "language_model.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
392
+ "language_model.model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
393
+ "language_model.model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
394
+ "language_model.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
395
+ "language_model.model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
396
+ "language_model.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
397
+ "language_model.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
398
+ "language_model.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
399
+ "language_model.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
400
+ "language_model.model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
401
+ "language_model.model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
402
+ "language_model.model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
403
+ "language_model.model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
404
+ "language_model.model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
405
+ "language_model.model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
406
+ "language_model.model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
407
+ "language_model.model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
408
+ "language_model.model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
409
+ "language_model.model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
410
+ "language_model.model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
411
+ "language_model.model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
412
+ "language_model.model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
413
+ "language_model.model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
414
+ "language_model.model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
415
+ "language_model.model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
416
+ "language_model.model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
417
+ "language_model.model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
418
+ "language_model.model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
419
+ "language_model.model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
420
+ "language_model.model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
421
+ "language_model.model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
422
+ "language_model.model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
423
+ "language_model.model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
424
+ "language_model.model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
425
+ "language_model.model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
426
+ "language_model.model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
427
+ "language_model.model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
428
+ "language_model.model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
429
+ "language_model.model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
430
+ "language_model.model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
431
+ "language_model.model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
432
+ "language_model.model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
433
+ "language_model.model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
434
+ "language_model.model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
435
+ "language_model.model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
436
+ "language_model.model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
437
+ "language_model.model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
438
+ "language_model.model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
439
+ "language_model.model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
440
+ "language_model.model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
441
+ "language_model.model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
442
+ "language_model.model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
443
+ "language_model.model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
444
+ "language_model.model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
445
+ "language_model.model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
446
+ "language_model.model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
447
+ "language_model.model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
448
+ "language_model.model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
449
+ "language_model.model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
450
+ "language_model.model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
451
+ "language_model.model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
452
+ "language_model.model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
453
+ "language_model.model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
454
+ "language_model.model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
455
+ "language_model.model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
456
+ "language_model.model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
457
+ "language_model.model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
458
+ "language_model.model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
459
+ "language_model.model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
460
+ "language_model.model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
461
+ "language_model.model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
462
+ "language_model.model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
463
+ "language_model.model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
464
+ "language_model.model.layers.19.input_layernorm.weight": "model-00003-of-00003.safetensors",
465
+ "language_model.model.layers.19.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
466
+ "language_model.model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
467
+ "language_model.model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
468
+ "language_model.model.layers.19.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
469
+ "language_model.model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
470
+ "language_model.model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
471
+ "language_model.model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
472
+ "language_model.model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
473
+ "language_model.model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
474
+ "language_model.model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
475
+ "language_model.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
476
+ "language_model.model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
477
+ "language_model.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
478
+ "language_model.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
479
+ "language_model.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
480
+ "language_model.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
481
+ "language_model.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
482
+ "language_model.model.layers.20.input_layernorm.weight": "model-00003-of-00003.safetensors",
483
+ "language_model.model.layers.20.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
484
+ "language_model.model.layers.20.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
485
+ "language_model.model.layers.20.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
486
+ "language_model.model.layers.20.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
487
+ "language_model.model.layers.20.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
488
+ "language_model.model.layers.20.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
489
+ "language_model.model.layers.20.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
490
+ "language_model.model.layers.20.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
491
+ "language_model.model.layers.21.input_layernorm.weight": "model-00003-of-00003.safetensors",
492
+ "language_model.model.layers.21.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
493
+ "language_model.model.layers.21.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
494
+ "language_model.model.layers.21.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
495
+ "language_model.model.layers.21.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
496
+ "language_model.model.layers.21.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
497
+ "language_model.model.layers.21.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
498
+ "language_model.model.layers.21.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
499
+ "language_model.model.layers.21.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
500
+ "language_model.model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
501
+ "language_model.model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
502
+ "language_model.model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
503
+ "language_model.model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
504
+ "language_model.model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
505
+ "language_model.model.layers.22.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
506
+ "language_model.model.layers.22.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
507
+ "language_model.model.layers.22.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
508
+ "language_model.model.layers.22.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
509
+ "language_model.model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
510
+ "language_model.model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
511
+ "language_model.model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
512
+ "language_model.model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
513
+ "language_model.model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
514
+ "language_model.model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
515
+ "language_model.model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
516
+ "language_model.model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
517
+ "language_model.model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
518
+ "language_model.model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
519
+ "language_model.model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
520
+ "language_model.model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
521
+ "language_model.model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
522
+ "language_model.model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
523
+ "language_model.model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
524
+ "language_model.model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
525
+ "language_model.model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
526
+ "language_model.model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
527
+ "language_model.model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
528
+ "language_model.model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
529
+ "language_model.model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
530
+ "language_model.model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
531
+ "language_model.model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
532
+ "language_model.model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
533
+ "language_model.model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
534
+ "language_model.model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
535
+ "language_model.model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
536
+ "language_model.model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
537
+ "language_model.model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
538
+ "language_model.model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
539
+ "language_model.model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
540
+ "language_model.model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
541
+ "language_model.model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
542
+ "language_model.model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
543
+ "language_model.model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
544
+ "language_model.model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
545
+ "language_model.model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
546
+ "language_model.model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
547
+ "language_model.model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
548
+ "language_model.model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
549
+ "language_model.model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
550
+ "language_model.model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
551
+ "language_model.model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
552
+ "language_model.model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
553
+ "language_model.model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
554
+ "language_model.model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
555
+ "language_model.model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
556
+ "language_model.model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
557
+ "language_model.model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
558
+ "language_model.model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
559
+ "language_model.model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
560
+ "language_model.model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
561
+ "language_model.model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
562
+ "language_model.model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
563
+ "language_model.model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
564
+ "language_model.model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
565
+ "language_model.model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
566
+ "language_model.model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
567
+ "language_model.model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
568
+ "language_model.model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
569
+ "language_model.model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
570
+ "language_model.model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
571
+ "language_model.model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
572
+ "language_model.model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
573
+ "language_model.model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
574
+ "language_model.model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
575
+ "language_model.model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
576
+ "language_model.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
577
+ "language_model.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
578
+ "language_model.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
579
+ "language_model.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
580
+ "language_model.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
581
+ "language_model.model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
582
+ "language_model.model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
583
+ "language_model.model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
584
+ "language_model.model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
585
+ "language_model.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
586
+ "language_model.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
587
+ "language_model.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
588
+ "language_model.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
589
+ "language_model.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
590
+ "language_model.model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
591
+ "language_model.model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
592
+ "language_model.model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
593
+ "language_model.model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
594
+ "language_model.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
595
+ "language_model.model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
596
+ "language_model.model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
597
+ "language_model.model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
598
+ "language_model.model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
599
+ "language_model.model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
600
+ "language_model.model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
601
+ "language_model.model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
602
+ "language_model.model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
603
+ "language_model.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
604
+ "language_model.model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
605
+ "language_model.model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
606
+ "language_model.model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
607
+ "language_model.model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
608
+ "language_model.model.layers.7.input_layernorm.weight": "model-00002-of-00003.safetensors",
609
+ "language_model.model.layers.7.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
610
+ "language_model.model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
611
+ "language_model.model.layers.7.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
612
+ "language_model.model.layers.7.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
613
+ "language_model.model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
614
+ "language_model.model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
615
+ "language_model.model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
616
+ "language_model.model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
617
+ "language_model.model.layers.8.input_layernorm.weight": "model-00002-of-00003.safetensors",
618
+ "language_model.model.layers.8.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
619
+ "language_model.model.layers.8.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
620
+ "language_model.model.layers.8.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
621
+ "language_model.model.layers.8.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
622
+ "language_model.model.layers.8.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
623
+ "language_model.model.layers.8.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
624
+ "language_model.model.layers.8.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
625
+ "language_model.model.layers.8.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
626
+ "language_model.model.layers.9.input_layernorm.weight": "model-00002-of-00003.safetensors",
627
+ "language_model.model.layers.9.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
628
+ "language_model.model.layers.9.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
629
+ "language_model.model.layers.9.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
630
+ "language_model.model.layers.9.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
631
+ "language_model.model.layers.9.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
632
+ "language_model.model.layers.9.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
633
+ "language_model.model.layers.9.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
634
+ "language_model.model.layers.9.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
635
+ "language_model.model.norm.weight": "model-00003-of-00003.safetensors",
636
+ "vision_model.vision_tower.attn_pool.kv.bias": "model-00001-of-00003.safetensors",
637
+ "vision_model.vision_tower.attn_pool.kv.weight": "model-00001-of-00003.safetensors",
638
+ "vision_model.vision_tower.attn_pool.latent": "model-00001-of-00003.safetensors",
639
+ "vision_model.vision_tower.attn_pool.mlp.fc1.bias": "model-00001-of-00003.safetensors",
640
+ "vision_model.vision_tower.attn_pool.mlp.fc1.weight": "model-00001-of-00003.safetensors",
641
+ "vision_model.vision_tower.attn_pool.mlp.fc2.bias": "model-00001-of-00003.safetensors",
642
+ "vision_model.vision_tower.attn_pool.mlp.fc2.weight": "model-00001-of-00003.safetensors",
643
+ "vision_model.vision_tower.attn_pool.norm.bias": "model-00001-of-00003.safetensors",
644
+ "vision_model.vision_tower.attn_pool.norm.weight": "model-00001-of-00003.safetensors",
645
+ "vision_model.vision_tower.attn_pool.proj.bias": "model-00001-of-00003.safetensors",
646
+ "vision_model.vision_tower.attn_pool.proj.weight": "model-00001-of-00003.safetensors",
647
+ "vision_model.vision_tower.attn_pool.q.bias": "model-00001-of-00003.safetensors",
648
+ "vision_model.vision_tower.attn_pool.q.weight": "model-00001-of-00003.safetensors",
649
+ "vision_model.vision_tower.blocks.0.attn.proj.bias": "model-00001-of-00003.safetensors",
650
+ "vision_model.vision_tower.blocks.0.attn.proj.weight": "model-00001-of-00003.safetensors",
651
+ "vision_model.vision_tower.blocks.0.attn.qkv.bias": "model-00001-of-00003.safetensors",
652
+ "vision_model.vision_tower.blocks.0.attn.qkv.weight": "model-00001-of-00003.safetensors",
653
+ "vision_model.vision_tower.blocks.0.mlp.fc1.bias": "model-00001-of-00003.safetensors",
654
+ "vision_model.vision_tower.blocks.0.mlp.fc1.weight": "model-00001-of-00003.safetensors",
655
+ "vision_model.vision_tower.blocks.0.mlp.fc2.bias": "model-00001-of-00003.safetensors",
656
+ "vision_model.vision_tower.blocks.0.mlp.fc2.weight": "model-00001-of-00003.safetensors",
657
+ "vision_model.vision_tower.blocks.0.norm1.bias": "model-00001-of-00003.safetensors",
658
+ "vision_model.vision_tower.blocks.0.norm1.weight": "model-00001-of-00003.safetensors",
659
+ "vision_model.vision_tower.blocks.0.norm2.bias": "model-00001-of-00003.safetensors",
660
+ "vision_model.vision_tower.blocks.0.norm2.weight": "model-00001-of-00003.safetensors",
661
+ "vision_model.vision_tower.blocks.1.attn.proj.bias": "model-00001-of-00003.safetensors",
662
+ "vision_model.vision_tower.blocks.1.attn.proj.weight": "model-00001-of-00003.safetensors",
663
+ "vision_model.vision_tower.blocks.1.attn.qkv.bias": "model-00001-of-00003.safetensors",
664
+ "vision_model.vision_tower.blocks.1.attn.qkv.weight": "model-00001-of-00003.safetensors",
665
+ "vision_model.vision_tower.blocks.1.mlp.fc1.bias": "model-00001-of-00003.safetensors",
666
+ "vision_model.vision_tower.blocks.1.mlp.fc1.weight": "model-00001-of-00003.safetensors",
667
+ "vision_model.vision_tower.blocks.1.mlp.fc2.bias": "model-00001-of-00003.safetensors",
668
+ "vision_model.vision_tower.blocks.1.mlp.fc2.weight": "model-00001-of-00003.safetensors",
669
+ "vision_model.vision_tower.blocks.1.norm1.bias": "model-00001-of-00003.safetensors",
670
+ "vision_model.vision_tower.blocks.1.norm1.weight": "model-00001-of-00003.safetensors",
671
+ "vision_model.vision_tower.blocks.1.norm2.bias": "model-00001-of-00003.safetensors",
672
+ "vision_model.vision_tower.blocks.1.norm2.weight": "model-00001-of-00003.safetensors",
673
+ "vision_model.vision_tower.blocks.10.attn.proj.bias": "model-00001-of-00003.safetensors",
674
+ "vision_model.vision_tower.blocks.10.attn.proj.weight": "model-00001-of-00003.safetensors",
675
+ "vision_model.vision_tower.blocks.10.attn.qkv.bias": "model-00001-of-00003.safetensors",
676
+ "vision_model.vision_tower.blocks.10.attn.qkv.weight": "model-00001-of-00003.safetensors",
677
+ "vision_model.vision_tower.blocks.10.mlp.fc1.bias": "model-00001-of-00003.safetensors",
678
+ "vision_model.vision_tower.blocks.10.mlp.fc1.weight": "model-00001-of-00003.safetensors",
679
+ "vision_model.vision_tower.blocks.10.mlp.fc2.bias": "model-00001-of-00003.safetensors",
680
+ "vision_model.vision_tower.blocks.10.mlp.fc2.weight": "model-00001-of-00003.safetensors",
681
+ "vision_model.vision_tower.blocks.10.norm1.bias": "model-00001-of-00003.safetensors",
682
+ "vision_model.vision_tower.blocks.10.norm1.weight": "model-00001-of-00003.safetensors",
683
+ "vision_model.vision_tower.blocks.10.norm2.bias": "model-00001-of-00003.safetensors",
684
+ "vision_model.vision_tower.blocks.10.norm2.weight": "model-00001-of-00003.safetensors",
685
+ "vision_model.vision_tower.blocks.11.attn.proj.bias": "model-00001-of-00003.safetensors",
686
+ "vision_model.vision_tower.blocks.11.attn.proj.weight": "model-00001-of-00003.safetensors",
687
+ "vision_model.vision_tower.blocks.11.attn.qkv.bias": "model-00001-of-00003.safetensors",
688
+ "vision_model.vision_tower.blocks.11.attn.qkv.weight": "model-00001-of-00003.safetensors",
689
+ "vision_model.vision_tower.blocks.11.mlp.fc1.bias": "model-00001-of-00003.safetensors",
690
+ "vision_model.vision_tower.blocks.11.mlp.fc1.weight": "model-00001-of-00003.safetensors",
691
+ "vision_model.vision_tower.blocks.11.mlp.fc2.bias": "model-00001-of-00003.safetensors",
692
+ "vision_model.vision_tower.blocks.11.mlp.fc2.weight": "model-00001-of-00003.safetensors",
693
+ "vision_model.vision_tower.blocks.11.norm1.bias": "model-00001-of-00003.safetensors",
694
+ "vision_model.vision_tower.blocks.11.norm1.weight": "model-00001-of-00003.safetensors",
695
+ "vision_model.vision_tower.blocks.11.norm2.bias": "model-00001-of-00003.safetensors",
696
+ "vision_model.vision_tower.blocks.11.norm2.weight": "model-00001-of-00003.safetensors",
697
+ "vision_model.vision_tower.blocks.12.attn.proj.bias": "model-00001-of-00003.safetensors",
698
+ "vision_model.vision_tower.blocks.12.attn.proj.weight": "model-00001-of-00003.safetensors",
699
+ "vision_model.vision_tower.blocks.12.attn.qkv.bias": "model-00001-of-00003.safetensors",
700
+ "vision_model.vision_tower.blocks.12.attn.qkv.weight": "model-00001-of-00003.safetensors",
701
+ "vision_model.vision_tower.blocks.12.mlp.fc1.bias": "model-00001-of-00003.safetensors",
702
+ "vision_model.vision_tower.blocks.12.mlp.fc1.weight": "model-00001-of-00003.safetensors",
703
+ "vision_model.vision_tower.blocks.12.mlp.fc2.bias": "model-00001-of-00003.safetensors",
704
+ "vision_model.vision_tower.blocks.12.mlp.fc2.weight": "model-00001-of-00003.safetensors",
705
+ "vision_model.vision_tower.blocks.12.norm1.bias": "model-00001-of-00003.safetensors",
706
+ "vision_model.vision_tower.blocks.12.norm1.weight": "model-00001-of-00003.safetensors",
707
+ "vision_model.vision_tower.blocks.12.norm2.bias": "model-00001-of-00003.safetensors",
708
+ "vision_model.vision_tower.blocks.12.norm2.weight": "model-00001-of-00003.safetensors",
709
+ "vision_model.vision_tower.blocks.13.attn.proj.bias": "model-00001-of-00003.safetensors",
710
+ "vision_model.vision_tower.blocks.13.attn.proj.weight": "model-00001-of-00003.safetensors",
711
+ "vision_model.vision_tower.blocks.13.attn.qkv.bias": "model-00001-of-00003.safetensors",
712
+ "vision_model.vision_tower.blocks.13.attn.qkv.weight": "model-00001-of-00003.safetensors",
713
+ "vision_model.vision_tower.blocks.13.mlp.fc1.bias": "model-00001-of-00003.safetensors",
714
+ "vision_model.vision_tower.blocks.13.mlp.fc1.weight": "model-00001-of-00003.safetensors",
715
+ "vision_model.vision_tower.blocks.13.mlp.fc2.bias": "model-00001-of-00003.safetensors",
716
+ "vision_model.vision_tower.blocks.13.mlp.fc2.weight": "model-00001-of-00003.safetensors",
717
+ "vision_model.vision_tower.blocks.13.norm1.bias": "model-00001-of-00003.safetensors",
718
+ "vision_model.vision_tower.blocks.13.norm1.weight": "model-00001-of-00003.safetensors",
719
+ "vision_model.vision_tower.blocks.13.norm2.bias": "model-00001-of-00003.safetensors",
720
+ "vision_model.vision_tower.blocks.13.norm2.weight": "model-00001-of-00003.safetensors",
721
+ "vision_model.vision_tower.blocks.14.attn.proj.bias": "model-00001-of-00003.safetensors",
722
+ "vision_model.vision_tower.blocks.14.attn.proj.weight": "model-00001-of-00003.safetensors",
723
+ "vision_model.vision_tower.blocks.14.attn.qkv.bias": "model-00001-of-00003.safetensors",
724
+ "vision_model.vision_tower.blocks.14.attn.qkv.weight": "model-00001-of-00003.safetensors",
725
+ "vision_model.vision_tower.blocks.14.mlp.fc1.bias": "model-00001-of-00003.safetensors",
726
+ "vision_model.vision_tower.blocks.14.mlp.fc1.weight": "model-00001-of-00003.safetensors",
727
+ "vision_model.vision_tower.blocks.14.mlp.fc2.bias": "model-00001-of-00003.safetensors",
728
+ "vision_model.vision_tower.blocks.14.mlp.fc2.weight": "model-00001-of-00003.safetensors",
729
+ "vision_model.vision_tower.blocks.14.norm1.bias": "model-00001-of-00003.safetensors",
730
+ "vision_model.vision_tower.blocks.14.norm1.weight": "model-00001-of-00003.safetensors",
731
+ "vision_model.vision_tower.blocks.14.norm2.bias": "model-00001-of-00003.safetensors",
732
+ "vision_model.vision_tower.blocks.14.norm2.weight": "model-00001-of-00003.safetensors",
733
+ "vision_model.vision_tower.blocks.15.attn.proj.bias": "model-00001-of-00003.safetensors",
734
+ "vision_model.vision_tower.blocks.15.attn.proj.weight": "model-00001-of-00003.safetensors",
735
+ "vision_model.vision_tower.blocks.15.attn.qkv.bias": "model-00001-of-00003.safetensors",
736
+ "vision_model.vision_tower.blocks.15.attn.qkv.weight": "model-00001-of-00003.safetensors",
737
+ "vision_model.vision_tower.blocks.15.mlp.fc1.bias": "model-00001-of-00003.safetensors",
738
+ "vision_model.vision_tower.blocks.15.mlp.fc1.weight": "model-00001-of-00003.safetensors",
739
+ "vision_model.vision_tower.blocks.15.mlp.fc2.bias": "model-00001-of-00003.safetensors",
740
+ "vision_model.vision_tower.blocks.15.mlp.fc2.weight": "model-00001-of-00003.safetensors",
741
+ "vision_model.vision_tower.blocks.15.norm1.bias": "model-00001-of-00003.safetensors",
742
+ "vision_model.vision_tower.blocks.15.norm1.weight": "model-00001-of-00003.safetensors",
743
+ "vision_model.vision_tower.blocks.15.norm2.bias": "model-00001-of-00003.safetensors",
744
+ "vision_model.vision_tower.blocks.15.norm2.weight": "model-00001-of-00003.safetensors",
745
+ "vision_model.vision_tower.blocks.16.attn.proj.bias": "model-00001-of-00003.safetensors",
746
+ "vision_model.vision_tower.blocks.16.attn.proj.weight": "model-00001-of-00003.safetensors",
747
+ "vision_model.vision_tower.blocks.16.attn.qkv.bias": "model-00001-of-00003.safetensors",
748
+ "vision_model.vision_tower.blocks.16.attn.qkv.weight": "model-00001-of-00003.safetensors",
749
+ "vision_model.vision_tower.blocks.16.mlp.fc1.bias": "model-00001-of-00003.safetensors",
750
+ "vision_model.vision_tower.blocks.16.mlp.fc1.weight": "model-00001-of-00003.safetensors",
751
+ "vision_model.vision_tower.blocks.16.mlp.fc2.bias": "model-00001-of-00003.safetensors",
752
+ "vision_model.vision_tower.blocks.16.mlp.fc2.weight": "model-00001-of-00003.safetensors",
753
+ "vision_model.vision_tower.blocks.16.norm1.bias": "model-00001-of-00003.safetensors",
754
+ "vision_model.vision_tower.blocks.16.norm1.weight": "model-00001-of-00003.safetensors",
755
+ "vision_model.vision_tower.blocks.16.norm2.bias": "model-00001-of-00003.safetensors",
756
+ "vision_model.vision_tower.blocks.16.norm2.weight": "model-00001-of-00003.safetensors",
757
+ "vision_model.vision_tower.blocks.17.attn.proj.bias": "model-00001-of-00003.safetensors",
758
+ "vision_model.vision_tower.blocks.17.attn.proj.weight": "model-00001-of-00003.safetensors",
759
+ "vision_model.vision_tower.blocks.17.attn.qkv.bias": "model-00001-of-00003.safetensors",
760
+ "vision_model.vision_tower.blocks.17.attn.qkv.weight": "model-00001-of-00003.safetensors",
761
+ "vision_model.vision_tower.blocks.17.mlp.fc1.bias": "model-00001-of-00003.safetensors",
762
+ "vision_model.vision_tower.blocks.17.mlp.fc1.weight": "model-00001-of-00003.safetensors",
763
+ "vision_model.vision_tower.blocks.17.mlp.fc2.bias": "model-00001-of-00003.safetensors",
764
+ "vision_model.vision_tower.blocks.17.mlp.fc2.weight": "model-00001-of-00003.safetensors",
765
+ "vision_model.vision_tower.blocks.17.norm1.bias": "model-00001-of-00003.safetensors",
766
+ "vision_model.vision_tower.blocks.17.norm1.weight": "model-00001-of-00003.safetensors",
767
+ "vision_model.vision_tower.blocks.17.norm2.bias": "model-00001-of-00003.safetensors",
768
+ "vision_model.vision_tower.blocks.17.norm2.weight": "model-00001-of-00003.safetensors",
769
+ "vision_model.vision_tower.blocks.18.attn.proj.bias": "model-00001-of-00003.safetensors",
770
+ "vision_model.vision_tower.blocks.18.attn.proj.weight": "model-00001-of-00003.safetensors",
771
+ "vision_model.vision_tower.blocks.18.attn.qkv.bias": "model-00001-of-00003.safetensors",
772
+ "vision_model.vision_tower.blocks.18.attn.qkv.weight": "model-00001-of-00003.safetensors",
773
+ "vision_model.vision_tower.blocks.18.mlp.fc1.bias": "model-00001-of-00003.safetensors",
774
+ "vision_model.vision_tower.blocks.18.mlp.fc1.weight": "model-00001-of-00003.safetensors",
775
+ "vision_model.vision_tower.blocks.18.mlp.fc2.bias": "model-00001-of-00003.safetensors",
776
+ "vision_model.vision_tower.blocks.18.mlp.fc2.weight": "model-00001-of-00003.safetensors",
777
+ "vision_model.vision_tower.blocks.18.norm1.bias": "model-00001-of-00003.safetensors",
778
+ "vision_model.vision_tower.blocks.18.norm1.weight": "model-00001-of-00003.safetensors",
779
+ "vision_model.vision_tower.blocks.18.norm2.bias": "model-00001-of-00003.safetensors",
780
+ "vision_model.vision_tower.blocks.18.norm2.weight": "model-00001-of-00003.safetensors",
781
+ "vision_model.vision_tower.blocks.19.attn.proj.bias": "model-00001-of-00003.safetensors",
782
+ "vision_model.vision_tower.blocks.19.attn.proj.weight": "model-00001-of-00003.safetensors",
783
+ "vision_model.vision_tower.blocks.19.attn.qkv.bias": "model-00001-of-00003.safetensors",
784
+ "vision_model.vision_tower.blocks.19.attn.qkv.weight": "model-00001-of-00003.safetensors",
785
+ "vision_model.vision_tower.blocks.19.mlp.fc1.bias": "model-00001-of-00003.safetensors",
786
+ "vision_model.vision_tower.blocks.19.mlp.fc1.weight": "model-00001-of-00003.safetensors",
787
+ "vision_model.vision_tower.blocks.19.mlp.fc2.bias": "model-00001-of-00003.safetensors",
788
+ "vision_model.vision_tower.blocks.19.mlp.fc2.weight": "model-00001-of-00003.safetensors",
789
+ "vision_model.vision_tower.blocks.19.norm1.bias": "model-00001-of-00003.safetensors",
790
+ "vision_model.vision_tower.blocks.19.norm1.weight": "model-00001-of-00003.safetensors",
791
+ "vision_model.vision_tower.blocks.19.norm2.bias": "model-00001-of-00003.safetensors",
792
+ "vision_model.vision_tower.blocks.19.norm2.weight": "model-00001-of-00003.safetensors",
793
+ "vision_model.vision_tower.blocks.2.attn.proj.bias": "model-00001-of-00003.safetensors",
794
+ "vision_model.vision_tower.blocks.2.attn.proj.weight": "model-00001-of-00003.safetensors",
795
+ "vision_model.vision_tower.blocks.2.attn.qkv.bias": "model-00001-of-00003.safetensors",
796
+ "vision_model.vision_tower.blocks.2.attn.qkv.weight": "model-00001-of-00003.safetensors",
797
+ "vision_model.vision_tower.blocks.2.mlp.fc1.bias": "model-00001-of-00003.safetensors",
798
+ "vision_model.vision_tower.blocks.2.mlp.fc1.weight": "model-00001-of-00003.safetensors",
799
+ "vision_model.vision_tower.blocks.2.mlp.fc2.bias": "model-00001-of-00003.safetensors",
800
+ "vision_model.vision_tower.blocks.2.mlp.fc2.weight": "model-00001-of-00003.safetensors",
801
+ "vision_model.vision_tower.blocks.2.norm1.bias": "model-00001-of-00003.safetensors",
802
+ "vision_model.vision_tower.blocks.2.norm1.weight": "model-00001-of-00003.safetensors",
803
+ "vision_model.vision_tower.blocks.2.norm2.bias": "model-00001-of-00003.safetensors",
804
+ "vision_model.vision_tower.blocks.2.norm2.weight": "model-00001-of-00003.safetensors",
805
+ "vision_model.vision_tower.blocks.20.attn.proj.bias": "model-00001-of-00003.safetensors",
806
+ "vision_model.vision_tower.blocks.20.attn.proj.weight": "model-00001-of-00003.safetensors",
807
+ "vision_model.vision_tower.blocks.20.attn.qkv.bias": "model-00001-of-00003.safetensors",
808
+ "vision_model.vision_tower.blocks.20.attn.qkv.weight": "model-00001-of-00003.safetensors",
809
+ "vision_model.vision_tower.blocks.20.mlp.fc1.bias": "model-00001-of-00003.safetensors",
810
+ "vision_model.vision_tower.blocks.20.mlp.fc1.weight": "model-00001-of-00003.safetensors",
811
+ "vision_model.vision_tower.blocks.20.mlp.fc2.bias": "model-00001-of-00003.safetensors",
812
+ "vision_model.vision_tower.blocks.20.mlp.fc2.weight": "model-00001-of-00003.safetensors",
813
+ "vision_model.vision_tower.blocks.20.norm1.bias": "model-00001-of-00003.safetensors",
814
+ "vision_model.vision_tower.blocks.20.norm1.weight": "model-00001-of-00003.safetensors",
815
+ "vision_model.vision_tower.blocks.20.norm2.bias": "model-00001-of-00003.safetensors",
816
+ "vision_model.vision_tower.blocks.20.norm2.weight": "model-00001-of-00003.safetensors",
817
+ "vision_model.vision_tower.blocks.21.attn.proj.bias": "model-00001-of-00003.safetensors",
818
+ "vision_model.vision_tower.blocks.21.attn.proj.weight": "model-00001-of-00003.safetensors",
819
+ "vision_model.vision_tower.blocks.21.attn.qkv.bias": "model-00001-of-00003.safetensors",
820
+ "vision_model.vision_tower.blocks.21.attn.qkv.weight": "model-00001-of-00003.safetensors",
821
+ "vision_model.vision_tower.blocks.21.mlp.fc1.bias": "model-00001-of-00003.safetensors",
822
+ "vision_model.vision_tower.blocks.21.mlp.fc1.weight": "model-00001-of-00003.safetensors",
823
+ "vision_model.vision_tower.blocks.21.mlp.fc2.bias": "model-00001-of-00003.safetensors",
824
+ "vision_model.vision_tower.blocks.21.mlp.fc2.weight": "model-00001-of-00003.safetensors",
825
+ "vision_model.vision_tower.blocks.21.norm1.bias": "model-00001-of-00003.safetensors",
826
+ "vision_model.vision_tower.blocks.21.norm1.weight": "model-00001-of-00003.safetensors",
827
+ "vision_model.vision_tower.blocks.21.norm2.bias": "model-00001-of-00003.safetensors",
828
+ "vision_model.vision_tower.blocks.21.norm2.weight": "model-00001-of-00003.safetensors",
829
+ "vision_model.vision_tower.blocks.22.attn.proj.bias": "model-00001-of-00003.safetensors",
830
+ "vision_model.vision_tower.blocks.22.attn.proj.weight": "model-00001-of-00003.safetensors",
831
+ "vision_model.vision_tower.blocks.22.attn.qkv.bias": "model-00001-of-00003.safetensors",
832
+ "vision_model.vision_tower.blocks.22.attn.qkv.weight": "model-00001-of-00003.safetensors",
833
+ "vision_model.vision_tower.blocks.22.mlp.fc1.bias": "model-00001-of-00003.safetensors",
834
+ "vision_model.vision_tower.blocks.22.mlp.fc1.weight": "model-00001-of-00003.safetensors",
835
+ "vision_model.vision_tower.blocks.22.mlp.fc2.bias": "model-00001-of-00003.safetensors",
836
+ "vision_model.vision_tower.blocks.22.mlp.fc2.weight": "model-00001-of-00003.safetensors",
837
+ "vision_model.vision_tower.blocks.22.norm1.bias": "model-00001-of-00003.safetensors",
838
+ "vision_model.vision_tower.blocks.22.norm1.weight": "model-00001-of-00003.safetensors",
839
+ "vision_model.vision_tower.blocks.22.norm2.bias": "model-00001-of-00003.safetensors",
840
+ "vision_model.vision_tower.blocks.22.norm2.weight": "model-00001-of-00003.safetensors",
841
+ "vision_model.vision_tower.blocks.23.attn.proj.bias": "model-00001-of-00003.safetensors",
842
+ "vision_model.vision_tower.blocks.23.attn.proj.weight": "model-00001-of-00003.safetensors",
843
+ "vision_model.vision_tower.blocks.23.attn.qkv.bias": "model-00001-of-00003.safetensors",
844
+ "vision_model.vision_tower.blocks.23.attn.qkv.weight": "model-00001-of-00003.safetensors",
845
+ "vision_model.vision_tower.blocks.23.mlp.fc1.bias": "model-00001-of-00003.safetensors",
846
+ "vision_model.vision_tower.blocks.23.mlp.fc1.weight": "model-00001-of-00003.safetensors",
847
+ "vision_model.vision_tower.blocks.23.mlp.fc2.bias": "model-00001-of-00003.safetensors",
848
+ "vision_model.vision_tower.blocks.23.mlp.fc2.weight": "model-00001-of-00003.safetensors",
849
+ "vision_model.vision_tower.blocks.23.norm1.bias": "model-00001-of-00003.safetensors",
850
+ "vision_model.vision_tower.blocks.23.norm1.weight": "model-00001-of-00003.safetensors",
851
+ "vision_model.vision_tower.blocks.23.norm2.bias": "model-00001-of-00003.safetensors",
852
+ "vision_model.vision_tower.blocks.23.norm2.weight": "model-00001-of-00003.safetensors",
853
+ "vision_model.vision_tower.blocks.3.attn.proj.bias": "model-00001-of-00003.safetensors",
854
+ "vision_model.vision_tower.blocks.3.attn.proj.weight": "model-00001-of-00003.safetensors",
855
+ "vision_model.vision_tower.blocks.3.attn.qkv.bias": "model-00001-of-00003.safetensors",
856
+ "vision_model.vision_tower.blocks.3.attn.qkv.weight": "model-00001-of-00003.safetensors",
857
+ "vision_model.vision_tower.blocks.3.mlp.fc1.bias": "model-00001-of-00003.safetensors",
858
+ "vision_model.vision_tower.blocks.3.mlp.fc1.weight": "model-00001-of-00003.safetensors",
859
+ "vision_model.vision_tower.blocks.3.mlp.fc2.bias": "model-00001-of-00003.safetensors",
860
+ "vision_model.vision_tower.blocks.3.mlp.fc2.weight": "model-00001-of-00003.safetensors",
861
+ "vision_model.vision_tower.blocks.3.norm1.bias": "model-00001-of-00003.safetensors",
862
+ "vision_model.vision_tower.blocks.3.norm1.weight": "model-00001-of-00003.safetensors",
863
+ "vision_model.vision_tower.blocks.3.norm2.bias": "model-00001-of-00003.safetensors",
864
+ "vision_model.vision_tower.blocks.3.norm2.weight": "model-00001-of-00003.safetensors",
865
+ "vision_model.vision_tower.blocks.4.attn.proj.bias": "model-00001-of-00003.safetensors",
866
+ "vision_model.vision_tower.blocks.4.attn.proj.weight": "model-00001-of-00003.safetensors",
867
+ "vision_model.vision_tower.blocks.4.attn.qkv.bias": "model-00001-of-00003.safetensors",
868
+ "vision_model.vision_tower.blocks.4.attn.qkv.weight": "model-00001-of-00003.safetensors",
869
+ "vision_model.vision_tower.blocks.4.mlp.fc1.bias": "model-00001-of-00003.safetensors",
870
+ "vision_model.vision_tower.blocks.4.mlp.fc1.weight": "model-00001-of-00003.safetensors",
871
+ "vision_model.vision_tower.blocks.4.mlp.fc2.bias": "model-00001-of-00003.safetensors",
872
+ "vision_model.vision_tower.blocks.4.mlp.fc2.weight": "model-00001-of-00003.safetensors",
873
+ "vision_model.vision_tower.blocks.4.norm1.bias": "model-00001-of-00003.safetensors",
874
+ "vision_model.vision_tower.blocks.4.norm1.weight": "model-00001-of-00003.safetensors",
875
+ "vision_model.vision_tower.blocks.4.norm2.bias": "model-00001-of-00003.safetensors",
876
+ "vision_model.vision_tower.blocks.4.norm2.weight": "model-00001-of-00003.safetensors",
877
+ "vision_model.vision_tower.blocks.5.attn.proj.bias": "model-00001-of-00003.safetensors",
878
+ "vision_model.vision_tower.blocks.5.attn.proj.weight": "model-00001-of-00003.safetensors",
879
+ "vision_model.vision_tower.blocks.5.attn.qkv.bias": "model-00001-of-00003.safetensors",
880
+ "vision_model.vision_tower.blocks.5.attn.qkv.weight": "model-00001-of-00003.safetensors",
881
+ "vision_model.vision_tower.blocks.5.mlp.fc1.bias": "model-00001-of-00003.safetensors",
882
+ "vision_model.vision_tower.blocks.5.mlp.fc1.weight": "model-00001-of-00003.safetensors",
883
+ "vision_model.vision_tower.blocks.5.mlp.fc2.bias": "model-00001-of-00003.safetensors",
884
+ "vision_model.vision_tower.blocks.5.mlp.fc2.weight": "model-00001-of-00003.safetensors",
885
+ "vision_model.vision_tower.blocks.5.norm1.bias": "model-00001-of-00003.safetensors",
886
+ "vision_model.vision_tower.blocks.5.norm1.weight": "model-00001-of-00003.safetensors",
887
+ "vision_model.vision_tower.blocks.5.norm2.bias": "model-00001-of-00003.safetensors",
888
+ "vision_model.vision_tower.blocks.5.norm2.weight": "model-00001-of-00003.safetensors",
889
+ "vision_model.vision_tower.blocks.6.attn.proj.bias": "model-00001-of-00003.safetensors",
890
+ "vision_model.vision_tower.blocks.6.attn.proj.weight": "model-00001-of-00003.safetensors",
891
+ "vision_model.vision_tower.blocks.6.attn.qkv.bias": "model-00001-of-00003.safetensors",
892
+ "vision_model.vision_tower.blocks.6.attn.qkv.weight": "model-00001-of-00003.safetensors",
893
+ "vision_model.vision_tower.blocks.6.mlp.fc1.bias": "model-00001-of-00003.safetensors",
894
+ "vision_model.vision_tower.blocks.6.mlp.fc1.weight": "model-00001-of-00003.safetensors",
895
+ "vision_model.vision_tower.blocks.6.mlp.fc2.bias": "model-00001-of-00003.safetensors",
896
+ "vision_model.vision_tower.blocks.6.mlp.fc2.weight": "model-00001-of-00003.safetensors",
897
+ "vision_model.vision_tower.blocks.6.norm1.bias": "model-00001-of-00003.safetensors",
898
+ "vision_model.vision_tower.blocks.6.norm1.weight": "model-00001-of-00003.safetensors",
899
+ "vision_model.vision_tower.blocks.6.norm2.bias": "model-00001-of-00003.safetensors",
900
+ "vision_model.vision_tower.blocks.6.norm2.weight": "model-00001-of-00003.safetensors",
901
+ "vision_model.vision_tower.blocks.7.attn.proj.bias": "model-00001-of-00003.safetensors",
902
+ "vision_model.vision_tower.blocks.7.attn.proj.weight": "model-00001-of-00003.safetensors",
903
+ "vision_model.vision_tower.blocks.7.attn.qkv.bias": "model-00001-of-00003.safetensors",
904
+ "vision_model.vision_tower.blocks.7.attn.qkv.weight": "model-00001-of-00003.safetensors",
905
+ "vision_model.vision_tower.blocks.7.mlp.fc1.bias": "model-00001-of-00003.safetensors",
906
+ "vision_model.vision_tower.blocks.7.mlp.fc1.weight": "model-00001-of-00003.safetensors",
907
+ "vision_model.vision_tower.blocks.7.mlp.fc2.bias": "model-00001-of-00003.safetensors",
908
+ "vision_model.vision_tower.blocks.7.mlp.fc2.weight": "model-00001-of-00003.safetensors",
909
+ "vision_model.vision_tower.blocks.7.norm1.bias": "model-00001-of-00003.safetensors",
910
+ "vision_model.vision_tower.blocks.7.norm1.weight": "model-00001-of-00003.safetensors",
911
+ "vision_model.vision_tower.blocks.7.norm2.bias": "model-00001-of-00003.safetensors",
912
+ "vision_model.vision_tower.blocks.7.norm2.weight": "model-00001-of-00003.safetensors",
913
+ "vision_model.vision_tower.blocks.8.attn.proj.bias": "model-00001-of-00003.safetensors",
914
+ "vision_model.vision_tower.blocks.8.attn.proj.weight": "model-00001-of-00003.safetensors",
915
+ "vision_model.vision_tower.blocks.8.attn.qkv.bias": "model-00001-of-00003.safetensors",
916
+ "vision_model.vision_tower.blocks.8.attn.qkv.weight": "model-00001-of-00003.safetensors",
917
+ "vision_model.vision_tower.blocks.8.mlp.fc1.bias": "model-00001-of-00003.safetensors",
918
+ "vision_model.vision_tower.blocks.8.mlp.fc1.weight": "model-00001-of-00003.safetensors",
919
+ "vision_model.vision_tower.blocks.8.mlp.fc2.bias": "model-00001-of-00003.safetensors",
920
+ "vision_model.vision_tower.blocks.8.mlp.fc2.weight": "model-00001-of-00003.safetensors",
921
+ "vision_model.vision_tower.blocks.8.norm1.bias": "model-00001-of-00003.safetensors",
922
+ "vision_model.vision_tower.blocks.8.norm1.weight": "model-00001-of-00003.safetensors",
923
+ "vision_model.vision_tower.blocks.8.norm2.bias": "model-00001-of-00003.safetensors",
924
+ "vision_model.vision_tower.blocks.8.norm2.weight": "model-00001-of-00003.safetensors",
925
+ "vision_model.vision_tower.blocks.9.attn.proj.bias": "model-00001-of-00003.safetensors",
926
+ "vision_model.vision_tower.blocks.9.attn.proj.weight": "model-00001-of-00003.safetensors",
927
+ "vision_model.vision_tower.blocks.9.attn.qkv.bias": "model-00001-of-00003.safetensors",
928
+ "vision_model.vision_tower.blocks.9.attn.qkv.weight": "model-00001-of-00003.safetensors",
929
+ "vision_model.vision_tower.blocks.9.mlp.fc1.bias": "model-00001-of-00003.safetensors",
930
+ "vision_model.vision_tower.blocks.9.mlp.fc1.weight": "model-00001-of-00003.safetensors",
931
+ "vision_model.vision_tower.blocks.9.mlp.fc2.bias": "model-00001-of-00003.safetensors",
932
+ "vision_model.vision_tower.blocks.9.mlp.fc2.weight": "model-00001-of-00003.safetensors",
933
+ "vision_model.vision_tower.blocks.9.norm1.bias": "model-00001-of-00003.safetensors",
934
+ "vision_model.vision_tower.blocks.9.norm1.weight": "model-00001-of-00003.safetensors",
935
+ "vision_model.vision_tower.blocks.9.norm2.bias": "model-00001-of-00003.safetensors",
936
+ "vision_model.vision_tower.blocks.9.norm2.weight": "model-00001-of-00003.safetensors",
937
+ "vision_model.vision_tower.norm.bias": "model-00001-of-00003.safetensors",
938
+ "vision_model.vision_tower.norm.weight": "model-00001-of-00003.safetensors",
939
+ "vision_model.vision_tower.patch_embed.proj.bias": "model-00001-of-00003.safetensors",
940
+ "vision_model.vision_tower.patch_embed.proj.weight": "model-00001-of-00003.safetensors",
941
+ "vision_model.vision_tower.pos_embed": "model-00001-of-00003.safetensors"
942
+ }
943
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "background_color": [
3
+ 127,
4
+ 127,
5
+ 127
6
+ ],
7
+ "do_normalize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "VLMImageProcessor",
14
+ "image_size": 384,
15
+ "image_std": [
16
+ 0.5,
17
+ 0.5,
18
+ 0.5
19
+ ],
20
+ "min_size": 14,
21
+ "processor_class": "VLChatProcessor",
22
+ "rescale_factor": 0.00392156862745098
23
+ }
processor_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_special_token": false,
3
+ "ignore_id": -100,
4
+ "image_end_tag": "<end_of_image>",
5
+ "image_start_tag": "<begin_of_image>",
6
+ "image_tag": "<image_placeholder>",
7
+ "mask_prompt": true,
8
+ "num_image_tokens": 576,
9
+ "pad_tag": "<\uff5c\u2581pad\u2581\uff5c>",
10
+ "processor_class": "VLChatProcessor",
11
+ "sft_format": "deepseek"
12
+ }
rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:746da0c0f7deb757d99392a8fa181dafca631d6a381f2af4877bd3847bb9d198
3
+ size 14960
rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d38b053fe5eb6ae5ccf9adf9bd1c8889a57b4ff88703fe6902d6d5de3712167d
3
+ size 14960
rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b24871a6ff9b65654aed42574fc40d257ad398da6636c16d9b4a7a7dff22c2a
3
+ size 14960
rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc2f8e6fbf67f520f63c85a37a3e247b7317e64ae74ec25d2270c4915438e9aa
3
+ size 14960
special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<image_placeholder>",
4
+ "<patch_placeholder>",
5
+ "<|ref|>",
6
+ "<|/ref|>",
7
+ "<|det|>",
8
+ "<|/det|>",
9
+ "<|grounding|>",
10
+ "<|User|>",
11
+ "<|Assistant|>"
12
+ ],
13
+ "bos_token": {
14
+ "content": "<｜begin▁of▁sentence｜>",
15
+ "lstrip": false,
16
+ "normalized": true,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "eos_token": {
21
+ "content": "<｜end▁of▁sentence｜>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "pad_token": {
28
+ "content": "<｜▁pad▁｜>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
@@ -0,0 +1,1733 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.11074197120708748,
5
+ "eval_steps": 500,
6
+ "global_step": 100,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "completion_length": 73.265625,
13
+ "epoch": 0.0011074197120708748,
14
+ "grad_norm": 0.5076314806938171,
15
+ "kl": 0.0,
16
+ "learning_rate": 9.99375e-07,
17
+ "loss": -0.018259915290400386,
18
+ "reward": 2.2648561000823975,
19
+ "reward_std": 0.32521533221006393,
20
+ "rewards/GDino": 0.84943026304245,
21
+ "rewards/GIT": 0.5776679813861847,
22
+ "rewards/HPSv2": 0.2639656066894531,
23
+ "rewards/ORM": 0.5737921893596649,
24
+ "self_certainty_semantic": -25.4375,
25
+ "self_certainty_token": -22.0,
26
+ "step": 1
27
+ },
28
+ {
29
+ "completion_length": 56.0,
30
+ "epoch": 0.0022148394241417496,
31
+ "grad_norm": 0.5364330410957336,
32
+ "kl": 0.001522064208984375,
33
+ "learning_rate": 9.9875e-07,
34
+ "loss": 0.00348748016403988,
35
+ "reward": 1.7680926322937012,
36
+ "reward_std": 0.41801488399505615,
37
+ "rewards/GDino": 0.6529064476490021,
38
+ "rewards/GIT": 0.19494981318712234,
39
+ "rewards/HPSv2": 0.24983596801757812,
40
+ "rewards/ORM": 0.6704004406929016,
41
+ "self_certainty_semantic": -25.4375,
42
+ "self_certainty_token": -21.0,
43
+ "step": 2
44
+ },
45
+ {
46
+ "completion_length": 55.4375,
47
+ "epoch": 0.0033222591362126247,
48
+ "grad_norm": 0.5614722967147827,
49
+ "kl": 0.001556396484375,
50
+ "learning_rate": 9.98125e-07,
51
+ "loss": 0.01565772108733654,
52
+ "reward": 1.6570448875427246,
53
+ "reward_std": 0.3965621292591095,
54
+ "rewards/GDino": 0.6382372081279755,
55
+ "rewards/GIT": 0.37795570492744446,
56
+ "rewards/HPSv2": 0.24709796905517578,
57
+ "rewards/ORM": 0.3937540017068386,
58
+ "self_certainty_semantic": -25.1875,
59
+ "self_certainty_token": -20.9375,
60
+ "step": 3
61
+ },
62
+ {
63
+ "completion_length": 65.34375,
64
+ "epoch": 0.004429678848283499,
65
+ "grad_norm": 2.5736770629882812,
66
+ "kl": 0.0016021728515625,
67
+ "learning_rate": 9.975e-07,
68
+ "loss": -0.0012893765233457088,
69
+ "reward": 2.061529755592346,
70
+ "reward_std": 0.4106704443693161,
71
+ "rewards/GDino": 0.7796730995178223,
72
+ "rewards/GIT": 0.43717896938323975,
73
+ "rewards/HPSv2": 0.24744796752929688,
74
+ "rewards/ORM": 0.5972296595573425,
75
+ "self_certainty_semantic": -25.5,
76
+ "self_certainty_token": -22.0,
77
+ "step": 4
78
+ },
79
+ {
80
+ "completion_length": 63.578125,
81
+ "epoch": 0.005537098560354375,
82
+ "grad_norm": 0.48238250613212585,
83
+ "kl": 0.001575469970703125,
84
+ "learning_rate": 9.968749999999999e-07,
85
+ "loss": 0.020129199139773846,
86
+ "reward": 1.5302643775939941,
87
+ "reward_std": 0.44902199506759644,
88
+ "rewards/GDino": 0.6246840953826904,
89
+ "rewards/GIT": 0.23608428239822388,
90
+ "rewards/HPSv2": 0.2453451156616211,
91
+ "rewards/ORM": 0.42415088415145874,
92
+ "self_certainty_semantic": -25.625,
93
+ "self_certainty_token": -22.1875,
94
+ "step": 5
95
+ },
96
+ {
97
+ "completion_length": 60.65625,
98
+ "epoch": 0.006644518272425249,
99
+ "grad_norm": 0.8221905827522278,
100
+ "kl": 0.001674652099609375,
101
+ "learning_rate": 9.9625e-07,
102
+ "loss": 0.0192068200558424,
103
+ "reward": 2.1602972745895386,
104
+ "reward_std": 0.23134037852287292,
105
+ "rewards/GDino": 0.783700168132782,
106
+ "rewards/GIT": 0.452057421207428,
107
+ "rewards/HPSv2": 0.274627685546875,
108
+ "rewards/ORM": 0.6499120593070984,
109
+ "self_certainty_semantic": -25.3125,
110
+ "self_certainty_token": -22.0,
111
+ "step": 6
112
+ },
113
+ {
114
+ "completion_length": 65.453125,
115
+ "epoch": 0.007751937984496124,
116
+ "grad_norm": 0.433403879404068,
117
+ "kl": 0.0016021728515625,
118
+ "learning_rate": 9.956249999999999e-07,
119
+ "loss": 0.028950304724276066,
120
+ "reward": 1.7097668647766113,
121
+ "reward_std": 0.5880981385707855,
122
+ "rewards/GDino": 0.5914062708616257,
123
+ "rewards/GIT": 0.15753822773694992,
124
+ "rewards/HPSv2": 0.25023555755615234,
125
+ "rewards/ORM": 0.7105867862701416,
126
+ "self_certainty_semantic": -25.5,
127
+ "self_certainty_token": -21.75,
128
+ "step": 7
129
+ },
130
+ {
131
+ "completion_length": 74.90625,
132
+ "epoch": 0.008859357696566999,
133
+ "grad_norm": 0.41245806217193604,
134
+ "kl": 0.00152587890625,
135
+ "learning_rate": 9.95e-07,
136
+ "loss": -0.016540683340281248,
137
+ "reward": 1.785366177558899,
138
+ "reward_std": 0.39637817442417145,
139
+ "rewards/GDino": 0.7011832594871521,
140
+ "rewards/GIT": 0.3848375529050827,
141
+ "rewards/HPSv2": 0.2445659637451172,
142
+ "rewards/ORM": 0.45477938652038574,
143
+ "self_certainty_semantic": -25.375,
144
+ "self_certainty_token": -20.875,
145
+ "step": 8
146
+ },
147
+ {
148
+ "completion_length": 61.828125,
149
+ "epoch": 0.009966777408637873,
150
+ "grad_norm": 0.3924250602722168,
151
+ "kl": 0.001617431640625,
152
+ "learning_rate": 9.94375e-07,
153
+ "loss": 0.03069412149488926,
154
+ "reward": 2.0813064575195312,
155
+ "reward_std": 0.5435488224029541,
156
+ "rewards/GDino": 0.736801415681839,
157
+ "rewards/GIT": 0.32275132089853287,
158
+ "rewards/HPSv2": 0.26233673095703125,
159
+ "rewards/ORM": 0.759416937828064,
160
+ "self_certainty_semantic": -25.5625,
161
+ "self_certainty_token": -21.1875,
162
+ "step": 9
163
+ },
164
+ {
165
+ "completion_length": 62.796875,
166
+ "epoch": 0.01107419712070875,
167
+ "grad_norm": 0.5886948704719543,
168
+ "kl": 0.00164031982421875,
169
+ "learning_rate": 9.9375e-07,
170
+ "loss": -0.009089878294616938,
171
+ "reward": 1.8167259693145752,
172
+ "reward_std": 0.4427160769701004,
173
+ "rewards/GDino": 0.6997816860675812,
174
+ "rewards/GIT": 0.4742187559604645,
175
+ "rewards/HPSv2": 0.2480792999267578,
176
+ "rewards/ORM": 0.3946462571620941,
177
+ "self_certainty_semantic": -25.5,
178
+ "self_certainty_token": -21.125,
179
+ "step": 10
180
+ },
181
+ {
182
+ "completion_length": 64.09375,
183
+ "epoch": 0.012181616832779624,
184
+ "grad_norm": 0.6388463377952576,
185
+ "kl": 0.0016326904296875,
186
+ "learning_rate": 9.93125e-07,
187
+ "loss": -0.011163983959704638,
188
+ "reward": 2.250586152076721,
189
+ "reward_std": 0.29546695202589035,
190
+ "rewards/GDino": 0.7932291626930237,
191
+ "rewards/GIT": 0.5437096580862999,
192
+ "rewards/HPSv2": 0.25614356994628906,
193
+ "rewards/ORM": 0.657503753900528,
194
+ "self_certainty_semantic": -25.25,
195
+ "self_certainty_token": -21.0625,
196
+ "step": 11
197
+ },
198
+ {
199
+ "completion_length": 73.265625,
200
+ "epoch": 0.013289036544850499,
201
+ "grad_norm": 0.37963175773620605,
202
+ "kl": 0.001583099365234375,
203
+ "learning_rate": 9.925e-07,
204
+ "loss": 0.009535952471196651,
205
+ "reward": 1.8723560571670532,
206
+ "reward_std": 0.48824670910835266,
207
+ "rewards/GDino": 0.671429455280304,
208
+ "rewards/GIT": 0.4155814051628113,
209
+ "rewards/HPSv2": 0.2387409210205078,
210
+ "rewards/ORM": 0.5466042459011078,
211
+ "self_certainty_semantic": -25.4375,
212
+ "self_certainty_token": -21.625,
213
+ "step": 12
214
+ },
215
+ {
216
+ "completion_length": 55.015625,
217
+ "epoch": 0.014396456256921373,
218
+ "grad_norm": 0.5844080448150635,
219
+ "kl": 0.001674652099609375,
220
+ "learning_rate": 9.91875e-07,
221
+ "loss": 0.0034986711107194424,
222
+ "reward": 1.7595484256744385,
223
+ "reward_std": 0.3697086051106453,
224
+ "rewards/GDino": 0.7100214958190918,
225
+ "rewards/GIT": 0.26869260519742966,
226
+ "rewards/HPSv2": 0.24958419799804688,
227
+ "rewards/ORM": 0.53125,
228
+ "self_certainty_semantic": -25.25,
229
+ "self_certainty_token": -21.375,
230
+ "step": 13
231
+ },
232
+ {
233
+ "completion_length": 55.65625,
234
+ "epoch": 0.015503875968992248,
235
+ "grad_norm": 0.5192797780036926,
236
+ "kl": 0.001674652099609375,
237
+ "learning_rate": 9.912499999999998e-07,
238
+ "loss": 0.010001872200518847,
239
+ "reward": 2.201015591621399,
240
+ "reward_std": 0.4899330288171768,
241
+ "rewards/GDino": 0.8140625059604645,
242
+ "rewards/GIT": 0.4328514188528061,
243
+ "rewards/HPSv2": 0.2431640625,
244
+ "rewards/ORM": 0.7109375,
245
+ "self_certainty_semantic": -25.4375,
246
+ "self_certainty_token": -22.0,
247
+ "step": 14
248
+ },
249
+ {
250
+ "completion_length": 64.0,
251
+ "epoch": 0.016611295681063124,
252
+ "grad_norm": 0.46844616532325745,
253
+ "kl": 0.00174713134765625,
254
+ "learning_rate": 9.90625e-07,
255
+ "loss": 0.0017675042618066072,
256
+ "reward": 2.433342456817627,
257
+ "reward_std": 0.33736473321914673,
258
+ "rewards/GDino": 0.9153576791286469,
259
+ "rewards/GIT": 0.5124611556529999,
260
+ "rewards/HPSv2": 0.2507901191711426,
261
+ "rewards/ORM": 0.7547334432601929,
262
+ "self_certainty_semantic": -25.4375,
263
+ "self_certainty_token": -21.6875,
264
+ "step": 15
265
+ },
266
+ {
267
+ "completion_length": 53.203125,
268
+ "epoch": 0.017718715393133997,
269
+ "grad_norm": 0.49579355120658875,
270
+ "kl": 0.001758575439453125,
271
+ "learning_rate": 9.9e-07,
272
+ "loss": 0.003856237977743149,
273
+ "reward": 1.6368815302848816,
274
+ "reward_std": 0.42226114869117737,
275
+ "rewards/GDino": 0.6432631015777588,
276
+ "rewards/GIT": 0.2906690910458565,
277
+ "rewards/HPSv2": 0.25169944763183594,
278
+ "rewards/ORM": 0.45124977827072144,
279
+ "self_certainty_semantic": -25.3125,
280
+ "self_certainty_token": -20.8125,
281
+ "step": 16
282
+ },
283
+ {
284
+ "completion_length": 76.28125,
285
+ "epoch": 0.018826135105204873,
286
+ "grad_norm": 0.5296036601066589,
287
+ "kl": 0.001590728759765625,
288
+ "learning_rate": 9.89375e-07,
289
+ "loss": -0.003345506265759468,
290
+ "reward": 1.7861530184745789,
291
+ "reward_std": 0.5057752877473831,
292
+ "rewards/GDino": 0.6293700635433197,
293
+ "rewards/GIT": 0.2197464406490326,
294
+ "rewards/HPSv2": 0.26516151428222656,
295
+ "rewards/ORM": 0.671875,
296
+ "self_certainty_semantic": -25.5,
297
+ "self_certainty_token": -20.6875,
298
+ "step": 17
299
+ },
300
+ {
301
+ "completion_length": 58.0625,
302
+ "epoch": 0.019933554817275746,
303
+ "grad_norm": 0.6577962636947632,
304
+ "kl": 0.00174713134765625,
305
+ "learning_rate": 9.8875e-07,
306
+ "loss": -0.019500677473843098,
307
+ "reward": 2.303292393684387,
308
+ "reward_std": 0.2609405145049095,
309
+ "rewards/GDino": 0.8339102566242218,
310
+ "rewards/GIT": 0.5853700041770935,
311
+ "rewards/HPSv2": 0.24338722229003906,
312
+ "rewards/ORM": 0.640625,
313
+ "self_certainty_semantic": -25.1875,
314
+ "self_certainty_token": -21.25,
315
+ "step": 18
316
+ },
317
+ {
318
+ "completion_length": 50.859375,
319
+ "epoch": 0.021040974529346623,
320
+ "grad_norm": 0.3543226718902588,
321
+ "kl": 0.00182342529296875,
322
+ "learning_rate": 9.88125e-07,
323
+ "loss": -0.00019507110118865967,
324
+ "reward": 1.6344053149223328,
325
+ "reward_std": 0.47374215722084045,
326
+ "rewards/GDino": 0.705148845911026,
327
+ "rewards/GIT": 0.2559727430343628,
328
+ "rewards/HPSv2": 0.2541370391845703,
329
+ "rewards/ORM": 0.41914665699005127,
330
+ "self_certainty_semantic": -25.25,
331
+ "self_certainty_token": -22.3125,
332
+ "step": 19
333
+ },
334
+ {
335
+ "completion_length": 65.921875,
336
+ "epoch": 0.0221483942414175,
337
+ "grad_norm": 0.5358290672302246,
338
+ "kl": 0.001781463623046875,
339
+ "learning_rate": 9.875e-07,
340
+ "loss": 0.007933363318443298,
341
+ "reward": 1.9504321217536926,
342
+ "reward_std": 0.3728322237730026,
343
+ "rewards/GDino": 0.6606760025024414,
344
+ "rewards/GIT": 0.48046815395355225,
345
+ "rewards/HPSv2": 0.24678802490234375,
346
+ "rewards/ORM": 0.5625,
347
+ "self_certainty_semantic": -25.125,
348
+ "self_certainty_token": -21.0625,
349
+ "step": 20
350
+ },
351
+ {
352
+ "completion_length": 59.3125,
353
+ "epoch": 0.023255813953488372,
354
+ "grad_norm": 2.0912797451019287,
355
+ "kl": 0.001811981201171875,
356
+ "learning_rate": 9.86875e-07,
357
+ "loss": -0.004398644436150789,
358
+ "reward": 2.252086877822876,
359
+ "reward_std": 0.44888848066329956,
360
+ "rewards/GDino": 0.798213005065918,
361
+ "rewards/GIT": 0.4853799045085907,
362
+ "rewards/HPSv2": 0.25956153869628906,
363
+ "rewards/ORM": 0.7089323997497559,
364
+ "self_certainty_semantic": -25.1875,
365
+ "self_certainty_token": -22.75,
366
+ "step": 21
367
+ },
368
+ {
369
+ "completion_length": 52.265625,
370
+ "epoch": 0.024363233665559248,
371
+ "grad_norm": 0.5790585875511169,
372
+ "kl": 0.00191497802734375,
373
+ "learning_rate": 9.862499999999999e-07,
374
+ "loss": 0.006876260507851839,
375
+ "reward": 1.9933909177780151,
376
+ "reward_std": 0.32367050647735596,
377
+ "rewards/GDino": 0.7134387493133545,
378
+ "rewards/GIT": 0.41087181866168976,
379
+ "rewards/HPSv2": 0.2721214294433594,
380
+ "rewards/ORM": 0.5969589203596115,
381
+ "self_certainty_semantic": -25.5,
382
+ "self_certainty_token": -22.375,
383
+ "step": 22
384
+ },
385
+ {
386
+ "completion_length": 59.375,
387
+ "epoch": 0.02547065337763012,
388
+ "grad_norm": 0.45692723989486694,
389
+ "kl": 0.001697540283203125,
390
+ "learning_rate": 9.85625e-07,
391
+ "loss": -0.00792664848268032,
392
+ "reward": 2.015365242958069,
393
+ "reward_std": 0.48256243765354156,
394
+ "rewards/GDino": 0.724082350730896,
395
+ "rewards/GIT": 0.42729710042476654,
396
+ "rewards/HPSv2": 0.2667560577392578,
397
+ "rewards/ORM": 0.5972296893596649,
398
+ "self_certainty_semantic": -25.375,
399
+ "self_certainty_token": -20.6875,
400
+ "step": 23
401
+ },
402
+ {
403
+ "completion_length": 55.203125,
404
+ "epoch": 0.026578073089700997,
405
+ "grad_norm": 0.46439889073371887,
406
+ "kl": 0.0016937255859375,
407
+ "learning_rate": 9.849999999999999e-07,
408
+ "loss": 0.0024933242239058018,
409
+ "reward": 2.460409939289093,
410
+ "reward_std": 0.4443647414445877,
411
+ "rewards/GDino": 0.8454739451408386,
412
+ "rewards/GIT": 0.6258784532546997,
413
+ "rewards/HPSv2": 0.2624950408935547,
414
+ "rewards/ORM": 0.7265625,
415
+ "self_certainty_semantic": -25.6875,
416
+ "self_certainty_token": -21.25,
417
+ "step": 24
418
+ },
419
+ {
420
+ "completion_length": 60.15625,
421
+ "epoch": 0.02768549280177187,
422
+ "grad_norm": 0.47176027297973633,
423
+ "kl": 0.001880645751953125,
424
+ "learning_rate": 9.84375e-07,
425
+ "loss": 0.005812188144773245,
426
+ "reward": 2.0174233317375183,
427
+ "reward_std": 0.40724658966064453,
428
+ "rewards/GDino": 0.7186038792133331,
429
+ "rewards/GIT": 0.4156235605478287,
430
+ "rewards/HPSv2": 0.26485633850097656,
431
+ "rewards/ORM": 0.6183395236730576,
432
+ "self_certainty_semantic": -25.5,
433
+ "self_certainty_token": -22.0,
434
+ "step": 25
435
+ },
436
+ {
437
+ "completion_length": 53.21875,
438
+ "epoch": 0.028792912513842746,
439
+ "grad_norm": 0.716375470161438,
440
+ "kl": 0.00209808349609375,
441
+ "learning_rate": 9.8375e-07,
442
+ "loss": 0.02397427149116993,
443
+ "reward": 2.186239182949066,
444
+ "reward_std": 0.46710920333862305,
445
+ "rewards/GDino": 0.7593750059604645,
446
+ "rewards/GIT": 0.5171153843402863,
447
+ "rewards/HPSv2": 0.2734565734863281,
448
+ "rewards/ORM": 0.6362921893596649,
449
+ "self_certainty_semantic": -25.25,
450
+ "self_certainty_token": -23.0,
451
+ "step": 26
452
+ },
453
+ {
454
+ "completion_length": 58.421875,
455
+ "epoch": 0.029900332225913623,
456
+ "grad_norm": 0.428893119096756,
457
+ "kl": 0.00171661376953125,
458
+ "learning_rate": 9.83125e-07,
459
+ "loss": -0.005866332910954952,
460
+ "reward": 1.9681838750839233,
461
+ "reward_std": 0.3645169883966446,
462
+ "rewards/GDino": 0.7666666209697723,
463
+ "rewards/GIT": 0.4486802965402603,
464
+ "rewards/HPSv2": 0.2419452667236328,
465
+ "rewards/ORM": 0.5108915567398071,
466
+ "self_certainty_semantic": -25.375,
467
+ "self_certainty_token": -21.4375,
468
+ "step": 27
469
+ },
470
+ {
471
+ "completion_length": 63.328125,
472
+ "epoch": 0.031007751937984496,
473
+ "grad_norm": 0.5334203243255615,
474
+ "kl": 0.002010345458984375,
475
+ "learning_rate": 9.825e-07,
476
+ "loss": 0.012586410157382488,
477
+ "reward": 1.4134111404418945,
478
+ "reward_std": 0.3155324012041092,
479
+ "rewards/GDino": 0.6005972325801849,
480
+ "rewards/GIT": 0.11092349141836166,
481
+ "rewards/HPSv2": 0.2596569061279297,
482
+ "rewards/ORM": 0.44223344326019287,
483
+ "self_certainty_semantic": -25.3125,
484
+ "self_certainty_token": -20.5625,
485
+ "step": 28
486
+ },
487
+ {
488
+ "completion_length": 56.25,
489
+ "epoch": 0.03211517165005537,
490
+ "grad_norm": 0.40832045674324036,
491
+ "kl": 0.001819610595703125,
492
+ "learning_rate": 9.81875e-07,
493
+ "loss": 0.010300841182470322,
494
+ "reward": 2.465680956840515,
495
+ "reward_std": 0.298002652823925,
496
+ "rewards/GDino": 0.862500011920929,
497
+ "rewards/GIT": 0.6107669174671173,
498
+ "rewards/HPSv2": 0.28375244140625,
499
+ "rewards/ORM": 0.7086615860462189,
500
+ "self_certainty_semantic": -25.375,
501
+ "self_certainty_token": -21.0625,
502
+ "step": 29
503
+ },
504
+ {
505
+ "completion_length": 54.953125,
506
+ "epoch": 0.03322259136212625,
507
+ "grad_norm": 0.4050670266151428,
508
+ "kl": 0.002025604248046875,
509
+ "learning_rate": 9.8125e-07,
510
+ "loss": -0.001845305785536766,
511
+ "reward": 2.476737856864929,
512
+ "reward_std": 0.3756887763738632,
513
+ "rewards/GDino": 0.8967152833938599,
514
+ "rewards/GIT": 0.551719531416893,
515
+ "rewards/HPSv2": 0.24522781372070312,
516
+ "rewards/ORM": 0.7830752730369568,
517
+ "self_certainty_semantic": -25.1875,
518
+ "self_certainty_token": -21.25,
519
+ "step": 30
520
+ },
521
+ {
522
+ "completion_length": 74.84375,
523
+ "epoch": 0.03433001107419712,
524
+ "grad_norm": 0.7089686393737793,
525
+ "kl": 0.001865386962890625,
526
+ "learning_rate": 9.806249999999998e-07,
527
+ "loss": 0.023707949556410313,
528
+ "reward": 1.831493854522705,
529
+ "reward_std": 0.37860143184661865,
530
+ "rewards/GDino": 0.6287499666213989,
531
+ "rewards/GIT": 0.3833145350217819,
532
+ "rewards/HPSv2": 0.2413043975830078,
533
+ "rewards/ORM": 0.578125,
534
+ "self_certainty_semantic": -25.4375,
535
+ "self_certainty_token": -20.9375,
536
+ "step": 31
537
+ },
538
+ {
539
+ "completion_length": 71.0625,
540
+ "epoch": 0.035437430786267994,
541
+ "grad_norm": 0.45204266905784607,
542
+ "kl": 0.00200653076171875,
543
+ "learning_rate": 9.8e-07,
544
+ "loss": 0.014695112593472004,
545
+ "reward": 1.5279032588005066,
546
+ "reward_std": 0.5042913109064102,
547
+ "rewards/GDino": 0.6702238023281097,
548
+ "rewards/GIT": 0.24817809462547302,
549
+ "rewards/HPSv2": 0.2356252670288086,
550
+ "rewards/ORM": 0.37387609481811523,
551
+ "self_certainty_semantic": -25.375,
552
+ "self_certainty_token": -22.5625,
553
+ "step": 32
554
+ },
555
+ {
556
+ "completion_length": 59.703125,
557
+ "epoch": 0.036544850498338874,
558
+ "grad_norm": 0.4359590411186218,
559
+ "kl": 0.00201416015625,
560
+ "learning_rate": 9.79375e-07,
561
+ "loss": 0.00610552029684186,
562
+ "reward": 2.3108657598495483,
563
+ "reward_std": 0.4415571391582489,
564
+ "rewards/GDino": 0.8515625,
565
+ "rewards/GIT": 0.6067334115505219,
566
+ "rewards/HPSv2": 0.22726917266845703,
567
+ "rewards/ORM": 0.6253007054328918,
568
+ "self_certainty_semantic": -25.375,
569
+ "self_certainty_token": -21.5,
570
+ "step": 33
571
+ },
572
+ {
573
+ "completion_length": 58.046875,
574
+ "epoch": 0.03765227021040975,
575
+ "grad_norm": 0.5853399038314819,
576
+ "kl": 0.002033233642578125,
577
+ "learning_rate": 9.7875e-07,
578
+ "loss": 0.023541483096778393,
579
+ "reward": 2.012690246105194,
580
+ "reward_std": 0.4660336524248123,
581
+ "rewards/GDino": 0.6989582777023315,
582
+ "rewards/GIT": 0.40700431168079376,
583
+ "rewards/HPSv2": 0.24774932861328125,
584
+ "rewards/ORM": 0.6589783728122711,
585
+ "self_certainty_semantic": -25.3125,
586
+ "self_certainty_token": -21.8125,
587
+ "step": 34
588
+ },
589
+ {
590
+ "completion_length": 56.90625,
591
+ "epoch": 0.03875968992248062,
592
+ "grad_norm": 0.3787715435028076,
593
+ "kl": 0.001888275146484375,
594
+ "learning_rate": 9.78125e-07,
595
+ "loss": 0.003942073322832584,
596
+ "reward": 2.452033281326294,
597
+ "reward_std": 0.3410096764564514,
598
+ "rewards/GDino": 0.8359375298023224,
599
+ "rewards/GIT": 0.567652553319931,
600
+ "rewards/HPSv2": 0.2418804168701172,
601
+ "rewards/ORM": 0.806562751531601,
602
+ "self_certainty_semantic": -25.3125,
603
+ "self_certainty_token": -21.3125,
604
+ "step": 35
605
+ },
606
+ {
607
+ "completion_length": 66.0,
608
+ "epoch": 0.03986710963455149,
609
+ "grad_norm": 0.5305721163749695,
610
+ "kl": 0.005157470703125,
611
+ "learning_rate": 9.775e-07,
612
+ "loss": -0.003781900042667985,
613
+ "reward": 1.8618011474609375,
614
+ "reward_std": 0.4120703786611557,
615
+ "rewards/GDino": 0.6453125476837158,
616
+ "rewards/GIT": 0.4281370937824249,
617
+ "rewards/HPSv2": 0.24621009826660156,
618
+ "rewards/ORM": 0.5421415567398071,
619
+ "self_certainty_semantic": -25.5625,
620
+ "self_certainty_token": -20.9375,
621
+ "step": 36
622
+ },
623
+ {
624
+ "completion_length": 51.40625,
625
+ "epoch": 0.04097452934662237,
626
+ "grad_norm": 0.46515390276908875,
627
+ "kl": 0.002716064453125,
628
+ "learning_rate": 9.76875e-07,
629
+ "loss": 0.006902199704200029,
630
+ "reward": 1.9485998153686523,
631
+ "reward_std": 0.42147715389728546,
632
+ "rewards/GDino": 0.6951449513435364,
633
+ "rewards/GIT": 0.31057579815387726,
634
+ "rewards/HPSv2": 0.26158714294433594,
635
+ "rewards/ORM": 0.681291937828064,
636
+ "self_certainty_semantic": -25.3125,
637
+ "self_certainty_token": -22.1875,
638
+ "step": 37
639
+ },
640
+ {
641
+ "completion_length": 71.03125,
642
+ "epoch": 0.042081949058693245,
643
+ "grad_norm": 0.951810896396637,
644
+ "kl": 0.00226593017578125,
645
+ "learning_rate": 9.7625e-07,
646
+ "loss": 0.03428783547133207,
647
+ "reward": 1.9112213850021362,
648
+ "reward_std": 0.30633312463760376,
649
+ "rewards/GDino": 0.7401995956897736,
650
+ "rewards/GIT": 0.30288365483283997,
651
+ "rewards/HPSv2": 0.2552833557128906,
652
+ "rewards/ORM": 0.6128547042608261,
653
+ "self_certainty_semantic": -25.375,
654
+ "self_certainty_token": -21.125,
655
+ "step": 38
656
+ },
657
+ {
658
+ "completion_length": 67.6875,
659
+ "epoch": 0.04318936877076412,
660
+ "grad_norm": 0.6357575058937073,
661
+ "kl": 0.01482391357421875,
662
+ "learning_rate": 9.756249999999999e-07,
663
+ "loss": 0.023865018505603075,
664
+ "reward": 2.345404624938965,
665
+ "reward_std": 0.31367097795009613,
666
+ "rewards/GDino": 0.8703815042972565,
667
+ "rewards/GIT": 0.4902418553829193,
668
+ "rewards/HPSv2": 0.26603126525878906,
669
+ "rewards/ORM": 0.71875,
670
+ "self_certainty_semantic": -25.5625,
671
+ "self_certainty_token": -21.75,
672
+ "step": 39
673
+ },
674
+ {
675
+ "completion_length": 59.671875,
676
+ "epoch": 0.044296788482835,
677
+ "grad_norm": 0.5422465801239014,
678
+ "kl": 0.00281524658203125,
679
+ "learning_rate": 9.75e-07,
680
+ "loss": -0.018710695207118988,
681
+ "reward": 2.222834825515747,
682
+ "reward_std": 0.42842796444892883,
683
+ "rewards/GDino": 0.8634105622768402,
684
+ "rewards/GIT": 0.40908148139715195,
685
+ "rewards/HPSv2": 0.27498817443847656,
686
+ "rewards/ORM": 0.6753546893596649,
687
+ "self_certainty_semantic": -25.3125,
688
+ "self_certainty_token": -21.125,
689
+ "step": 40
690
+ },
691
+ {
692
+ "completion_length": 60.9375,
693
+ "epoch": 0.04540420819490587,
694
+ "grad_norm": 0.7511593103408813,
695
+ "kl": 0.00299072265625,
696
+ "learning_rate": 9.743749999999999e-07,
697
+ "loss": 0.005782268475741148,
698
+ "reward": 1.8980144262313843,
699
+ "reward_std": 0.3208035007119179,
700
+ "rewards/GDino": 0.6784752607345581,
701
+ "rewards/GIT": 0.3914954513311386,
702
+ "rewards/HPSv2": 0.24643898010253906,
703
+ "rewards/ORM": 0.5816046595573425,
704
+ "self_certainty_semantic": -25.375,
705
+ "self_certainty_token": -21.25,
706
+ "step": 41
707
+ },
708
+ {
709
+ "completion_length": 48.4375,
710
+ "epoch": 0.046511627906976744,
711
+ "grad_norm": 0.5177002549171448,
712
+ "kl": 0.0025177001953125,
713
+ "learning_rate": 9.7375e-07,
714
+ "loss": 0.045526545494794846,
715
+ "reward": 2.269711136817932,
716
+ "reward_std": 0.48014624416828156,
717
+ "rewards/GDino": 0.8855312466621399,
718
+ "rewards/GIT": 0.4437972754240036,
719
+ "rewards/HPSv2": 0.2572154998779297,
720
+ "rewards/ORM": 0.6831671893596649,
721
+ "self_certainty_semantic": -25.3125,
722
+ "self_certainty_token": -20.5625,
723
+ "step": 42
724
+ },
725
+ {
726
+ "completion_length": 67.875,
727
+ "epoch": 0.047619047619047616,
728
+ "grad_norm": 0.5885121822357178,
729
+ "kl": 0.002044677734375,
730
+ "learning_rate": 9.73125e-07,
731
+ "loss": 0.013573684729635715,
732
+ "reward": 1.6382005214691162,
733
+ "reward_std": 0.38919302821159363,
734
+ "rewards/GDino": 0.6114583313465118,
735
+ "rewards/GIT": 0.3806646466255188,
736
+ "rewards/HPSv2": 0.23286819458007812,
737
+ "rewards/ORM": 0.41320937871932983,
738
+ "self_certainty_semantic": -25.375,
739
+ "self_certainty_token": -21.0625,
740
+ "step": 43
741
+ },
742
+ {
743
+ "completion_length": 54.4375,
744
+ "epoch": 0.048726467331118496,
745
+ "grad_norm": 0.40727919340133667,
746
+ "kl": 0.0020751953125,
747
+ "learning_rate": 9.725e-07,
748
+ "loss": -0.01244093757122755,
749
+ "reward": 2.8831005096435547,
750
+ "reward_std": 0.31665875762701035,
751
+ "rewards/GDino": 0.9588541388511658,
752
+ "rewards/GIT": 0.7738310992717743,
753
+ "rewards/HPSv2": 0.2601909637451172,
754
+ "rewards/ORM": 0.8902243673801422,
755
+ "self_certainty_semantic": -25.5625,
756
+ "self_certainty_token": -21.0625,
757
+ "step": 44
758
+ },
759
+ {
760
+ "completion_length": 54.90625,
761
+ "epoch": 0.04983388704318937,
762
+ "grad_norm": 0.4928445816040039,
763
+ "kl": 0.0024566650390625,
764
+ "learning_rate": 9.71875e-07,
765
+ "loss": 0.00010553281754255295,
766
+ "reward": 2.4343937635421753,
767
+ "reward_std": 0.5984751731157303,
768
+ "rewards/GDino": 0.862500011920929,
769
+ "rewards/GIT": 0.5139474421739578,
770
+ "rewards/HPSv2": 0.26379966735839844,
771
+ "rewards/ORM": 0.7941466569900513,
772
+ "self_certainty_semantic": -25.5625,
773
+ "self_certainty_token": -21.25,
774
+ "step": 45
775
+ },
776
+ {
777
+ "completion_length": 56.078125,
778
+ "epoch": 0.05094130675526024,
779
+ "grad_norm": 0.37051326036453247,
780
+ "kl": 0.00231170654296875,
781
+ "learning_rate": 9.712499999999998e-07,
782
+ "loss": 0.007893505971878767,
783
+ "reward": 1.9575175046920776,
784
+ "reward_std": 0.3945648521184921,
785
+ "rewards/GDino": 0.5999999940395355,
786
+ "rewards/GIT": 0.32395021617412567,
787
+ "rewards/HPSv2": 0.26719093322753906,
788
+ "rewards/ORM": 0.7663763463497162,
789
+ "self_certainty_semantic": -25.1875,
790
+ "self_certainty_token": -22.0625,
791
+ "step": 46
792
+ },
793
+ {
794
+ "completion_length": 55.171875,
795
+ "epoch": 0.05204872646733112,
796
+ "grad_norm": 0.8945181369781494,
797
+ "kl": 0.0025634765625,
798
+ "learning_rate": 9.70625e-07,
799
+ "loss": -0.0013387980870902538,
800
+ "reward": 1.836871862411499,
801
+ "reward_std": 0.23468619585037231,
802
+ "rewards/GDino": 0.7209739089012146,
803
+ "rewards/GIT": 0.22856376320123672,
804
+ "rewards/HPSv2": 0.27921295166015625,
805
+ "rewards/ORM": 0.608121246099472,
806
+ "self_certainty_semantic": -25.3125,
807
+ "self_certainty_token": -21.6875,
808
+ "step": 47
809
+ },
810
+ {
811
+ "completion_length": 57.984375,
812
+ "epoch": 0.053156146179401995,
813
+ "grad_norm": 1.6689982414245605,
814
+ "kl": 0.00267791748046875,
815
+ "learning_rate": 9.7e-07,
816
+ "loss": 0.022647732868790627,
817
+ "reward": 1.454766035079956,
818
+ "reward_std": 0.40884387493133545,
819
+ "rewards/GDino": 0.6050891876220703,
820
+ "rewards/GIT": 0.0,
821
+ "rewards/HPSv2": 0.2698974609375,
822
+ "rewards/ORM": 0.5797793865203857,
823
+ "self_certainty_semantic": -25.3125,
824
+ "self_certainty_token": -21.625,
825
+ "step": 48
826
+ },
827
+ {
828
+ "completion_length": 58.046875,
829
+ "epoch": 0.05426356589147287,
830
+ "grad_norm": 0.4761441648006439,
831
+ "kl": 0.002048492431640625,
832
+ "learning_rate": 9.69375e-07,
833
+ "loss": 0.016307475278154016,
834
+ "reward": 1.9066129326820374,
835
+ "reward_std": 0.5319462567567825,
836
+ "rewards/GDino": 0.7744874656200409,
837
+ "rewards/GIT": 0.2370736114680767,
838
+ "rewards/HPSv2": 0.2514495849609375,
839
+ "rewards/ORM": 0.6436022371053696,
840
+ "self_certainty_semantic": -25.25,
841
+ "self_certainty_token": -21.625,
842
+ "step": 49
843
+ },
844
+ {
845
+ "completion_length": 61.0,
846
+ "epoch": 0.05537098560354374,
847
+ "grad_norm": 0.8074173331260681,
848
+ "kl": 0.0040283203125,
849
+ "learning_rate": 9.6875e-07,
850
+ "loss": 0.005913220578804612,
851
+ "reward": 2.0915883779525757,
852
+ "reward_std": 0.5395111739635468,
853
+ "rewards/GDino": 0.7859093248844147,
854
+ "rewards/GIT": 0.3929952085018158,
855
+ "rewards/HPSv2": 0.25482940673828125,
856
+ "rewards/ORM": 0.657854437828064,
857
+ "self_certainty_semantic": -25.5625,
858
+ "self_certainty_token": -22.25,
859
+ "step": 50
860
+ },
861
+ {
862
+ "completion_length": 44.359375,
863
+ "epoch": 0.05647840531561462,
864
+ "grad_norm": 0.5618427991867065,
865
+ "kl": 0.002471923828125,
866
+ "learning_rate": 9.68125e-07,
867
+ "loss": -0.003945098840631545,
868
+ "reward": 1.8058671951293945,
869
+ "reward_std": 0.5712144523859024,
870
+ "rewards/GDino": 0.7815796732902527,
871
+ "rewards/GIT": 0.2604931816458702,
872
+ "rewards/HPSv2": 0.27115440368652344,
873
+ "rewards/ORM": 0.49263995885849,
874
+ "self_certainty_semantic": -25.4375,
875
+ "self_certainty_token": -22.625,
876
+ "step": 51
877
+ },
878
+ {
879
+ "completion_length": 48.0,
880
+ "epoch": 0.05758582502768549,
881
+ "grad_norm": 107.57159423828125,
882
+ "kl": 26.37615966796875,
883
+ "learning_rate": 9.675e-07,
884
+ "loss": 0.27801212295889854,
885
+ "reward": 2.4165316820144653,
886
+ "reward_std": 0.2998274937272072,
887
+ "rewards/GDino": 0.9244791567325592,
888
+ "rewards/GIT": 0.6574473828077316,
889
+ "rewards/HPSv2": 0.2756366729736328,
890
+ "rewards/ORM": 0.5589684545993805,
891
+ "self_certainty_semantic": -25.4375,
892
+ "self_certainty_token": -21.375,
893
+ "step": 52
894
+ },
895
+ {
896
+ "completion_length": 52.140625,
897
+ "epoch": 0.058693244739756366,
898
+ "grad_norm": 0.4408358931541443,
899
+ "kl": 0.00232696533203125,
900
+ "learning_rate": 9.66875e-07,
901
+ "loss": 0.013528472045436502,
902
+ "reward": 1.8899905681610107,
903
+ "reward_std": 0.4558149725198746,
904
+ "rewards/GDino": 0.730059951543808,
905
+ "rewards/GIT": 0.39098620414733887,
906
+ "rewards/HPSv2": 0.24242782592773438,
907
+ "rewards/ORM": 0.5265165567398071,
908
+ "self_certainty_semantic": -25.0,
909
+ "self_certainty_token": -20.5625,
910
+ "step": 53
911
+ },
912
+ {
913
+ "completion_length": 58.390625,
914
+ "epoch": 0.059800664451827246,
915
+ "grad_norm": 0.48384228348731995,
916
+ "kl": 0.00225067138671875,
917
+ "learning_rate": 9.6625e-07,
918
+ "loss": 0.005568797350861132,
919
+ "reward": 1.638724684715271,
920
+ "reward_std": 0.41337575018405914,
921
+ "rewards/GDino": 0.6137361526489258,
922
+ "rewards/GIT": 0.24863167852163315,
923
+ "rewards/HPSv2": 0.24831581115722656,
924
+ "rewards/ORM": 0.5280410945415497,
925
+ "self_certainty_semantic": -25.4375,
926
+ "self_certainty_token": -20.9375,
927
+ "step": 54
928
+ },
929
+ {
930
+ "completion_length": 50.234375,
931
+ "epoch": 0.06090808416389812,
932
+ "grad_norm": 0.46963369846343994,
933
+ "kl": 0.0026397705078125,
934
+ "learning_rate": 9.65625e-07,
935
+ "loss": 0.009267964400351048,
936
+ "reward": 1.7191376686096191,
937
+ "reward_std": 0.521537572145462,
938
+ "rewards/GDino": 0.7086881995201111,
939
+ "rewards/GIT": 0.3270767852663994,
940
+ "rewards/HPSv2": 0.2678356170654297,
941
+ "rewards/ORM": 0.4155370891094208,
942
+ "self_certainty_semantic": -25.5,
943
+ "self_certainty_token": -21.6875,
944
+ "step": 55
945
+ },
946
+ {
947
+ "completion_length": 59.953125,
948
+ "epoch": 0.06201550387596899,
949
+ "grad_norm": 0.6913841366767883,
950
+ "kl": 0.0024261474609375,
951
+ "learning_rate": 9.649999999999999e-07,
952
+ "loss": 0.03414425998926163,
953
+ "reward": 1.9336698055267334,
954
+ "reward_std": 0.45749759674072266,
955
+ "rewards/GDino": 0.6963726580142975,
956
+ "rewards/GIT": 0.38425514101982117,
957
+ "rewards/HPSv2": 0.2471466064453125,
958
+ "rewards/ORM": 0.6058953106403351,
959
+ "self_certainty_semantic": -25.0625,
960
+ "self_certainty_token": -21.875,
961
+ "step": 56
962
+ },
963
+ {
964
+ "completion_length": 50.765625,
965
+ "epoch": 0.06312292358803986,
966
+ "grad_norm": 0.5066769123077393,
967
+ "kl": 0.002532958984375,
968
+ "learning_rate": 9.64375e-07,
969
+ "loss": 0.009842937346547842,
970
+ "reward": 1.8338811993598938,
971
+ "reward_std": 0.3951306492090225,
972
+ "rewards/GDino": 0.7909577786922455,
973
+ "rewards/GIT": 0.24781160056591034,
974
+ "rewards/HPSv2": 0.2509651184082031,
975
+ "rewards/ORM": 0.5441466569900513,
976
+ "self_certainty_semantic": -25.5,
977
+ "self_certainty_token": -21.4375,
978
+ "step": 57
979
+ },
980
+ {
981
+ "completion_length": 52.9375,
982
+ "epoch": 0.06423034330011074,
983
+ "grad_norm": 0.37791869044303894,
984
+ "kl": 0.002685546875,
985
+ "learning_rate": 9.637499999999999e-07,
986
+ "loss": 0.024126023054122925,
987
+ "reward": 1.8852884769439697,
988
+ "reward_std": 0.46756890416145325,
989
+ "rewards/GDino": 0.732811689376831,
990
+ "rewards/GIT": 0.38145140558481216,
991
+ "rewards/HPSv2": 0.2541465759277344,
992
+ "rewards/ORM": 0.5168787688016891,
993
+ "self_certainty_semantic": -25.4375,
994
+ "self_certainty_token": -21.25,
995
+ "step": 58
996
+ },
997
+ {
998
+ "completion_length": 47.234375,
999
+ "epoch": 0.06533776301218161,
1000
+ "grad_norm": 0.7410405278205872,
1001
+ "kl": 0.0026092529296875,
1002
+ "learning_rate": 9.63125e-07,
1003
+ "loss": -0.01674468442797661,
1004
+ "reward": 2.3462648391723633,
1005
+ "reward_std": 0.2433818019926548,
1006
+ "rewards/GDino": 0.8425607979297638,
1007
+ "rewards/GIT": 0.46571947634220123,
1008
+ "rewards/HPSv2": 0.2664222717285156,
1009
+ "rewards/ORM": 0.771562248468399,
1010
+ "self_certainty_semantic": -25.375,
1011
+ "self_certainty_token": -21.25,
1012
+ "step": 59
1013
+ },
1014
+ {
1015
+ "completion_length": 45.015625,
1016
+ "epoch": 0.0664451827242525,
1017
+ "grad_norm": 0.5326105952262878,
1018
+ "kl": 0.0026397705078125,
1019
+ "learning_rate": 9.624999999999999e-07,
1020
+ "loss": 0.003804182168096304,
1021
+ "reward": 2.036432147026062,
1022
+ "reward_std": 0.3990803211927414,
1023
+ "rewards/GDino": 0.8798050284385681,
1024
+ "rewards/GIT": 0.4744318723678589,
1025
+ "rewards/HPSv2": 0.238006591796875,
1026
+ "rewards/ORM": 0.44418865442276,
1027
+ "self_certainty_semantic": -25.3125,
1028
+ "self_certainty_token": -20.6875,
1029
+ "step": 60
1030
+ },
1031
+ {
1032
+ "completion_length": 65.0,
1033
+ "epoch": 0.06755260243632337,
1034
+ "grad_norm": 0.5713196396827698,
1035
+ "kl": 0.00235748291015625,
1036
+ "learning_rate": 9.61875e-07,
1037
+ "loss": 0.04368375800549984,
1038
+ "reward": 2.1398236751556396,
1039
+ "reward_std": 0.3530130609869957,
1040
+ "rewards/GDino": 0.7138020694255829,
1041
+ "rewards/GIT": 0.644903838634491,
1042
+ "rewards/HPSv2": 0.2529468536376953,
1043
+ "rewards/ORM": 0.5281709432601929,
1044
+ "self_certainty_semantic": -25.25,
1045
+ "self_certainty_token": -20.6875,
1046
+ "step": 61
1047
+ },
1048
+ {
1049
+ "completion_length": 54.9375,
1050
+ "epoch": 0.06866002214839424,
1051
+ "grad_norm": 5.612445831298828,
1052
+ "kl": 0.00339508056640625,
1053
+ "learning_rate": 9.6125e-07,
1054
+ "loss": 0.008875304833054543,
1055
+ "reward": 2.497900605201721,
1056
+ "reward_std": 0.41675496101379395,
1057
+ "rewards/GDino": 0.872697502374649,
1058
+ "rewards/GIT": 0.601748138666153,
1059
+ "rewards/HPSv2": 0.2640380859375,
1060
+ "rewards/ORM": 0.759416937828064,
1061
+ "self_certainty_semantic": -25.3125,
1062
+ "self_certainty_token": -20.875,
1063
+ "step": 62
1064
+ },
1065
+ {
1066
+ "completion_length": 49.734375,
1067
+ "epoch": 0.06976744186046512,
1068
+ "grad_norm": 0.5861217379570007,
1069
+ "kl": 0.003326416015625,
1070
+ "learning_rate": 9.606249999999998e-07,
1071
+ "loss": 0.01025251136161387,
1072
+ "reward": 2.2640050053596497,
1073
+ "reward_std": 0.48744213581085205,
1074
+ "rewards/GDino": 0.8172852694988251,
1075
+ "rewards/GIT": 0.44742196798324585,
1076
+ "rewards/HPSv2": 0.2430896759033203,
1077
+ "rewards/ORM": 0.7562080323696136,
1078
+ "self_certainty_semantic": -25.125,
1079
+ "self_certainty_token": -22.1875,
1080
+ "step": 63
1081
+ },
1082
+ {
1083
+ "completion_length": 64.375,
1084
+ "epoch": 0.07087486157253599,
1085
+ "grad_norm": 0.39266109466552734,
1086
+ "kl": 0.00298309326171875,
1087
+ "learning_rate": 9.6e-07,
1088
+ "loss": -0.005469436291605234,
1089
+ "reward": 1.6910768747329712,
1090
+ "reward_std": 0.2151722088456154,
1091
+ "rewards/GDino": 0.7097718715667725,
1092
+ "rewards/GIT": 0.32366037368774414,
1093
+ "rewards/HPSv2": 0.2576026916503906,
1094
+ "rewards/ORM": 0.40004195272922516,
1095
+ "self_certainty_semantic": -25.5,
1096
+ "self_certainty_token": -21.9375,
1097
+ "step": 64
1098
+ },
1099
+ {
1100
+ "completion_length": 61.515625,
1101
+ "epoch": 0.07198228128460686,
1102
+ "grad_norm": 0.705937922000885,
1103
+ "kl": 0.002685546875,
1104
+ "learning_rate": 9.59375e-07,
1105
+ "loss": 0.010601098649203777,
1106
+ "reward": 2.128853142261505,
1107
+ "reward_std": 0.4351096749305725,
1108
+ "rewards/GDino": 0.7197916805744171,
1109
+ "rewards/GIT": 0.6168824732303619,
1110
+ "rewards/HPSv2": 0.23163414001464844,
1111
+ "rewards/ORM": 0.5605448335409164,
1112
+ "self_certainty_semantic": -25.25,
1113
+ "self_certainty_token": -22.75,
1114
+ "step": 65
1115
+ },
1116
+ {
1117
+ "completion_length": 49.0,
1118
+ "epoch": 0.07308970099667775,
1119
+ "grad_norm": 0.4427480101585388,
1120
+ "kl": 0.002899169921875,
1121
+ "learning_rate": 9.5875e-07,
1122
+ "loss": 0.02646360918879509,
1123
+ "reward": 2.1654986143112183,
1124
+ "reward_std": 0.37753987312316895,
1125
+ "rewards/GDino": 0.6895833611488342,
1126
+ "rewards/GIT": 0.48387444019317627,
1127
+ "rewards/HPSv2": 0.2579364776611328,
1128
+ "rewards/ORM": 0.7341042160987854,
1129
+ "self_certainty_semantic": -25.25,
1130
+ "self_certainty_token": -21.5625,
1131
+ "step": 66
1132
+ },
1133
+ {
1134
+ "completion_length": 63.140625,
1135
+ "epoch": 0.07419712070874862,
1136
+ "grad_norm": 0.7619237899780273,
1137
+ "kl": 0.00284576416015625,
1138
+ "learning_rate": 9.58125e-07,
1139
+ "loss": 0.026691121514886618,
1140
+ "reward": 2.3450592160224915,
1141
+ "reward_std": 0.2740027904510498,
1142
+ "rewards/GDino": 0.8025760054588318,
1143
+ "rewards/GIT": 0.5677543580532074,
1144
+ "rewards/HPSv2": 0.2594585418701172,
1145
+ "rewards/ORM": 0.7152703106403351,
1146
+ "self_certainty_semantic": -25.25,
1147
+ "self_certainty_token": -21.875,
1148
+ "step": 67
1149
+ },
1150
+ {
1151
+ "completion_length": 50.171875,
1152
+ "epoch": 0.0753045404208195,
1153
+ "grad_norm": 0.4760603904724121,
1154
+ "kl": 0.0030517578125,
1155
+ "learning_rate": 9.575e-07,
1156
+ "loss": 0.022392848506569862,
1157
+ "reward": 1.6361079216003418,
1158
+ "reward_std": 0.33574268221855164,
1159
+ "rewards/GDino": 0.6061920523643494,
1160
+ "rewards/GIT": 0.31722745299339294,
1161
+ "rewards/HPSv2": 0.2595634460449219,
1162
+ "rewards/ORM": 0.453125,
1163
+ "self_certainty_semantic": -25.0625,
1164
+ "self_certainty_token": -21.5,
1165
+ "step": 68
1166
+ },
1167
+ {
1168
+ "completion_length": 55.046875,
1169
+ "epoch": 0.07641196013289037,
1170
+ "grad_norm": 0.5907943248748779,
1171
+ "kl": 0.00336456298828125,
1172
+ "learning_rate": 9.56875e-07,
1173
+ "loss": -0.0030646873638033867,
1174
+ "reward": 2.119426429271698,
1175
+ "reward_std": 0.298831045627594,
1176
+ "rewards/GDino": 0.8028125166893005,
1177
+ "rewards/GIT": 0.3893257826566696,
1178
+ "rewards/HPSv2": 0.2710380554199219,
1179
+ "rewards/ORM": 0.65625,
1180
+ "self_certainty_semantic": -25.4375,
1181
+ "self_certainty_token": -21.4375,
1182
+ "step": 69
1183
+ },
1184
+ {
1185
+ "completion_length": 51.140625,
1186
+ "epoch": 0.07751937984496124,
1187
+ "grad_norm": 0.47751373052597046,
1188
+ "kl": 0.00359344482421875,
1189
+ "learning_rate": 9.5625e-07,
1190
+ "loss": -0.011344656813889742,
1191
+ "reward": 1.4646188020706177,
1192
+ "reward_std": 0.5817874372005463,
1193
+ "rewards/GDino": 0.5935695767402649,
1194
+ "rewards/GIT": 0.23897356167435646,
1195
+ "rewards/HPSv2": 0.25234222412109375,
1196
+ "rewards/ORM": 0.37973344326019287,
1197
+ "self_certainty_semantic": -24.875,
1198
+ "self_certainty_token": -20.9375,
1199
+ "step": 70
1200
+ },
1201
+ {
1202
+ "completion_length": 55.609375,
1203
+ "epoch": 0.07862679955703211,
1204
+ "grad_norm": 0.5281980633735657,
1205
+ "kl": 0.0030364990234375,
1206
+ "learning_rate": 9.556249999999999e-07,
1207
+ "loss": -0.023217559792101383,
1208
+ "reward": 1.856022596359253,
1209
+ "reward_std": 0.4435942769050598,
1210
+ "rewards/GDino": 0.6947268545627594,
1211
+ "rewards/GIT": 0.28702250868082047,
1212
+ "rewards/HPSv2": 0.26489830017089844,
1213
+ "rewards/ORM": 0.609375,
1214
+ "self_certainty_semantic": -25.375,
1215
+ "self_certainty_token": -20.5,
1216
+ "step": 71
1217
+ },
1218
+ {
1219
+ "completion_length": 42.390625,
1220
+ "epoch": 0.07973421926910298,
1221
+ "grad_norm": 0.4538242518901825,
1222
+ "kl": 0.002960205078125,
1223
+ "learning_rate": 9.55e-07,
1224
+ "loss": 0.016265914775431156,
1225
+ "reward": 1.911847174167633,
1226
+ "reward_std": 0.4016146659851074,
1227
+ "rewards/GDino": 0.6731474995613098,
1228
+ "rewards/GIT": 0.46439771354198456,
1229
+ "rewards/HPSv2": 0.2497406005859375,
1230
+ "rewards/ORM": 0.524561420083046,
1231
+ "self_certainty_semantic": -25.0,
1232
+ "self_certainty_token": -20.875,
1233
+ "step": 72
1234
+ },
1235
+ {
1236
+ "completion_length": 53.28125,
1237
+ "epoch": 0.08084163898117387,
1238
+ "grad_norm": 0.5773823261260986,
1239
+ "kl": 0.00330352783203125,
1240
+ "learning_rate": 9.543749999999999e-07,
1241
+ "loss": -0.0016377167776226997,
1242
+ "reward": 2.114488363265991,
1243
+ "reward_std": 0.44427454471588135,
1244
+ "rewards/GDino": 0.8240922689437866,
1245
+ "rewards/GIT": 0.4950668513774872,
1246
+ "rewards/HPSv2": 0.24412155151367188,
1247
+ "rewards/ORM": 0.5512077808380127,
1248
+ "self_certainty_semantic": -25.375,
1249
+ "self_certainty_token": -21.0,
1250
+ "step": 73
1251
+ },
1252
+ {
1253
+ "completion_length": 56.296875,
1254
+ "epoch": 0.08194905869324474,
1255
+ "grad_norm": 0.43449509143829346,
1256
+ "kl": 0.0035247802734375,
1257
+ "learning_rate": 9.5375e-07,
1258
+ "loss": 0.03005522396415472,
1259
+ "reward": 2.32301664352417,
1260
+ "reward_std": 0.22542773187160492,
1261
+ "rewards/GDino": 0.864062488079071,
1262
+ "rewards/GIT": 0.5282620340585709,
1263
+ "rewards/HPSv2": 0.25408363342285156,
1264
+ "rewards/ORM": 0.6766084730625153,
1265
+ "self_certainty_semantic": -25.25,
1266
+ "self_certainty_token": -22.25,
1267
+ "step": 74
1268
+ },
1269
+ {
1270
+ "completion_length": 67.6875,
1271
+ "epoch": 0.08305647840531562,
1272
+ "grad_norm": 0.4218258857727051,
1273
+ "kl": 0.0028228759765625,
1274
+ "learning_rate": 9.53125e-07,
1275
+ "loss": 0.015081442426890135,
1276
+ "reward": 1.7625158429145813,
1277
+ "reward_std": 0.4334114193916321,
1278
+ "rewards/GDino": 0.6663236618041992,
1279
+ "rewards/GIT": 0.26877461373806,
1280
+ "rewards/HPSv2": 0.2647876739501953,
1281
+ "rewards/ORM": 0.5626298785209656,
1282
+ "self_certainty_semantic": -25.1875,
1283
+ "self_certainty_token": -21.125,
1284
+ "step": 75
1285
+ },
1286
+ {
1287
+ "completion_length": 62.15625,
1288
+ "epoch": 0.08416389811738649,
1289
+ "grad_norm": 0.45278123021125793,
1290
+ "kl": 0.00312042236328125,
1291
+ "learning_rate": 9.525e-07,
1292
+ "loss": 0.01650754688307643,
1293
+ "reward": 2.2938578128814697,
1294
+ "reward_std": 0.5077499151229858,
1295
+ "rewards/GDino": 0.7734375,
1296
+ "rewards/GIT": 0.6401466727256775,
1297
+ "rewards/HPSv2": 0.2568778991699219,
1298
+ "rewards/ORM": 0.6233955323696136,
1299
+ "self_certainty_semantic": -25.4375,
1300
+ "self_certainty_token": -21.125,
1301
+ "step": 76
1302
+ },
1303
+ {
1304
+ "completion_length": 50.984375,
1305
+ "epoch": 0.08527131782945736,
1306
+ "grad_norm": 0.5513558387756348,
1307
+ "kl": 0.004730224609375,
1308
+ "learning_rate": 9.51875e-07,
1309
+ "loss": -0.008258584188297391,
1310
+ "reward": 1.6354877948760986,
1311
+ "reward_std": 0.5420883148908615,
1312
+ "rewards/GDino": 0.643737405538559,
1313
+ "rewards/GIT": 0.20579323172569275,
1314
+ "rewards/HPSv2": 0.2405567169189453,
1315
+ "rewards/ORM": 0.5454003810882568,
1316
+ "self_certainty_semantic": -25.0625,
1317
+ "self_certainty_token": -22.1875,
1318
+ "step": 77
1319
+ },
1320
+ {
1321
+ "completion_length": 56.390625,
1322
+ "epoch": 0.08637873754152824,
1323
+ "grad_norm": 0.9578920602798462,
1324
+ "kl": 0.00360107421875,
1325
+ "learning_rate": 9.5125e-07,
1326
+ "loss": 0.0016261041164398193,
1327
+ "reward": 2.061507523059845,
1328
+ "reward_std": 0.2758500352501869,
1329
+ "rewards/GDino": 0.7561410367488861,
1330
+ "rewards/GIT": 0.33666322380304337,
1331
+ "rewards/HPSv2": 0.2762489318847656,
1332
+ "rewards/ORM": 0.6924542784690857,
1333
+ "self_certainty_semantic": -25.3125,
1334
+ "self_certainty_token": -21.1875,
1335
+ "step": 78
1336
+ },
1337
+ {
1338
+ "completion_length": 57.375,
1339
+ "epoch": 0.08748615725359911,
1340
+ "grad_norm": 0.46459418535232544,
1341
+ "kl": 0.004241943359375,
1342
+ "learning_rate": 9.50625e-07,
1343
+ "loss": -0.019409675151109695,
1344
+ "reward": 2.298323154449463,
1345
+ "reward_std": 0.22066934406757355,
1346
+ "rewards/GDino": 0.8136925399303436,
1347
+ "rewards/GIT": 0.6333461850881577,
1348
+ "rewards/HPSv2": 0.27008056640625,
1349
+ "rewards/ORM": 0.5812040567398071,
1350
+ "self_certainty_semantic": -25.3125,
1351
+ "self_certainty_token": -21.875,
1352
+ "step": 79
1353
+ },
1354
+ {
1355
+ "completion_length": 60.5625,
1356
+ "epoch": 0.08859357696567,
1357
+ "grad_norm": 0.4274587631225586,
1358
+ "kl": 0.004058837890625,
1359
+ "learning_rate": 9.499999999999999e-07,
1360
+ "loss": 0.013256619684398174,
1361
+ "reward": 1.6786987781524658,
1362
+ "reward_std": 0.3984425514936447,
1363
+ "rewards/GDino": 0.6007516384124756,
1364
+ "rewards/GIT": 0.18326736986637115,
1365
+ "rewards/HPSv2": 0.2720355987548828,
1366
+ "rewards/ORM": 0.6226442158222198,
1367
+ "self_certainty_semantic": -25.5,
1368
+ "self_certainty_token": -21.3125,
1369
+ "step": 80
1370
+ },
1371
+ {
1372
+ "completion_length": 58.03125,
1373
+ "epoch": 0.08970099667774087,
1374
+ "grad_norm": 0.9172859191894531,
1375
+ "kl": 0.00426483154296875,
1376
+ "learning_rate": 9.493749999999999e-07,
1377
+ "loss": 0.003496276680380106,
1378
+ "reward": 2.106017231941223,
1379
+ "reward_std": 0.30050399899482727,
1380
+ "rewards/GDino": 0.7440759837627411,
1381
+ "rewards/GIT": 0.3581302911043167,
1382
+ "rewards/HPSv2": 0.27126121520996094,
1383
+ "rewards/ORM": 0.7325496971607208,
1384
+ "self_certainty_semantic": -25.3125,
1385
+ "self_certainty_token": -20.875,
1386
+ "step": 81
1387
+ },
1388
+ {
1389
+ "completion_length": 49.5625,
1390
+ "epoch": 0.09080841638981174,
1391
+ "grad_norm": 0.4841405153274536,
1392
+ "kl": 0.00412750244140625,
1393
+ "learning_rate": 9.487499999999999e-07,
1394
+ "loss": 0.025506282225251198,
1395
+ "reward": 1.6879253387451172,
1396
+ "reward_std": 0.42353254556655884,
1397
+ "rewards/GDino": 0.6098452508449554,
1398
+ "rewards/GIT": 0.38033944368362427,
1399
+ "rewards/HPSv2": 0.2658271789550781,
1400
+ "rewards/ORM": 0.43191343545913696,
1401
+ "self_certainty_semantic": -25.4375,
1402
+ "self_certainty_token": -20.9375,
1403
+ "step": 82
1404
+ },
1405
+ {
1406
+ "completion_length": 48.328125,
1407
+ "epoch": 0.09191583610188261,
1408
+ "grad_norm": 0.492243230342865,
1409
+ "kl": 0.00345611572265625,
1410
+ "learning_rate": 9.481249999999999e-07,
1411
+ "loss": -0.0034960508346557617,
1412
+ "reward": 2.1111596822738647,
1413
+ "reward_std": 0.41540510952472687,
1414
+ "rewards/GDino": 0.7717877924442291,
1415
+ "rewards/GIT": 0.4860316216945648,
1416
+ "rewards/HPSv2": 0.2670021057128906,
1417
+ "rewards/ORM": 0.5863381326198578,
1418
+ "self_certainty_semantic": -25.25,
1419
+ "self_certainty_token": -21.4375,
1420
+ "step": 83
1421
+ },
1422
+ {
1423
+ "completion_length": 66.3125,
1424
+ "epoch": 0.09302325581395349,
1425
+ "grad_norm": 0.5617808699607849,
1426
+ "kl": 0.004180908203125,
1427
+ "learning_rate": 9.474999999999999e-07,
1428
+ "loss": 0.003248518332839012,
1429
+ "reward": 2.094790816307068,
1430
+ "reward_std": 0.3879907354712486,
1431
+ "rewards/GDino": 0.7973622679710388,
1432
+ "rewards/GIT": 0.632976621389389,
1433
+ "rewards/HPSv2": 0.24137306213378906,
1434
+ "rewards/ORM": 0.4230788052082062,
1435
+ "self_certainty_semantic": -25.25,
1436
+ "self_certainty_token": -21.1875,
1437
+ "step": 84
1438
+ },
1439
+ {
1440
+ "completion_length": 51.40625,
1441
+ "epoch": 0.09413067552602436,
1442
+ "grad_norm": 0.5695884823799133,
1443
+ "kl": 0.003204345703125,
1444
+ "learning_rate": 9.468749999999999e-07,
1445
+ "loss": 0.012543351389467716,
1446
+ "reward": 1.8675293326377869,
1447
+ "reward_std": 0.4282868355512619,
1448
+ "rewards/GDino": 0.6550000011920929,
1449
+ "rewards/GIT": 0.33260630816221237,
1450
+ "rewards/HPSv2": 0.24515533447265625,
1451
+ "rewards/ORM": 0.6347676515579224,
1452
+ "self_certainty_semantic": -25.0625,
1453
+ "self_certainty_token": -21.375,
1454
+ "step": 85
1455
+ },
1456
+ {
1457
+ "completion_length": 48.296875,
1458
+ "epoch": 0.09523809523809523,
1459
+ "grad_norm": 0.46590158343315125,
1460
+ "kl": 0.00469970703125,
1461
+ "learning_rate": 9.462499999999999e-07,
1462
+ "loss": 0.00347991194576025,
1463
+ "reward": 2.2731298208236694,
1464
+ "reward_std": 0.383390873670578,
1465
+ "rewards/GDino": 0.8246111273765564,
1466
+ "rewards/GIT": 0.33447980135679245,
1467
+ "rewards/HPSv2": 0.29212188720703125,
1468
+ "rewards/ORM": 0.821916937828064,
1469
+ "self_certainty_semantic": -25.375,
1470
+ "self_certainty_token": -22.0,
1471
+ "step": 86
1472
+ },
1473
+ {
1474
+ "completion_length": 54.390625,
1475
+ "epoch": 0.09634551495016612,
1476
+ "grad_norm": 0.5397853255271912,
1477
+ "kl": 0.004425048828125,
1478
+ "learning_rate": 9.45625e-07,
1479
+ "loss": 0.008617566898465157,
1480
+ "reward": 2.2459940314292908,
1481
+ "reward_std": 0.4676859378814697,
1482
+ "rewards/GDino": 0.7356771230697632,
1483
+ "rewards/GIT": 0.46453191339969635,
1484
+ "rewards/HPSv2": 0.26766395568847656,
1485
+ "rewards/ORM": 0.7781210243701935,
1486
+ "self_certainty_semantic": -25.25,
1487
+ "self_certainty_token": -21.0,
1488
+ "step": 87
1489
+ },
1490
+ {
1491
+ "completion_length": 42.15625,
1492
+ "epoch": 0.09745293466223699,
1493
+ "grad_norm": 0.48280662298202515,
1494
+ "kl": 0.00406646728515625,
1495
+ "learning_rate": 9.45e-07,
1496
+ "loss": 0.016791983507573605,
1497
+ "reward": 2.1528985500335693,
1498
+ "reward_std": 0.44025059044361115,
1499
+ "rewards/GDino": 0.7985424101352692,
1500
+ "rewards/GIT": 0.47699007391929626,
1501
+ "rewards/HPSv2": 0.2789325714111328,
1502
+ "rewards/ORM": 0.5984334945678711,
1503
+ "self_certainty_semantic": -25.375,
1504
+ "self_certainty_token": -20.875,
1505
+ "step": 88
1506
+ },
1507
+ {
1508
+ "completion_length": 48.25,
1509
+ "epoch": 0.09856035437430787,
1510
+ "grad_norm": 0.4512772560119629,
1511
+ "kl": 0.00405120849609375,
1512
+ "learning_rate": 9.44375e-07,
1513
+ "loss": -0.009609811007976532,
1514
+ "reward": 2.155352771282196,
1515
+ "reward_std": 0.3193782642483711,
1516
+ "rewards/GDino": 0.7525902688503265,
1517
+ "rewards/GIT": 0.4481022357940674,
1518
+ "rewards/HPSv2": 0.2619743347167969,
1519
+ "rewards/ORM": 0.6926859021186829,
1520
+ "self_certainty_semantic": -25.375,
1521
+ "self_certainty_token": -21.125,
1522
+ "step": 89
1523
+ },
1524
+ {
1525
+ "completion_length": 47.78125,
1526
+ "epoch": 0.09966777408637874,
1527
+ "grad_norm": 0.5204576849937439,
1528
+ "kl": 0.004425048828125,
1529
+ "learning_rate": 9.4375e-07,
1530
+ "loss": -0.017570611089468002,
1531
+ "reward": 2.3318194150924683,
1532
+ "reward_std": 0.3641355484724045,
1533
+ "rewards/GDino": 0.854687511920929,
1534
+ "rewards/GIT": 0.6271218061447144,
1535
+ "rewards/HPSv2": 0.26286888122558594,
1536
+ "rewards/ORM": 0.5871412754058838,
1537
+ "self_certainty_semantic": -25.5,
1538
+ "self_certainty_token": -21.4375,
1539
+ "step": 90
1540
+ },
1541
+ {
1542
+ "completion_length": 54.9375,
1543
+ "epoch": 0.10077519379844961,
1544
+ "grad_norm": 0.7515896558761597,
1545
+ "kl": 0.0042266845703125,
1546
+ "learning_rate": 9.43125e-07,
1547
+ "loss": 0.022024651989340782,
1548
+ "reward": 1.7255874276161194,
1549
+ "reward_std": 0.3924099802970886,
1550
+ "rewards/GDino": 0.6800954043865204,
1551
+ "rewards/GIT": 0.41760827600955963,
1552
+ "rewards/HPSv2": 0.22957611083984375,
1553
+ "rewards/ORM": 0.3983076214790344,
1554
+ "self_certainty_semantic": -24.875,
1555
+ "self_certainty_token": -21.125,
1556
+ "step": 91
1557
+ },
1558
+ {
1559
+ "completion_length": 51.90625,
1560
+ "epoch": 0.10188261351052048,
1561
+ "grad_norm": 0.6844750046730042,
1562
+ "kl": 0.004730224609375,
1563
+ "learning_rate": 9.425e-07,
1564
+ "loss": 0.017017286270856857,
1565
+ "reward": 1.7472361326217651,
1566
+ "reward_std": 0.49342362582683563,
1567
+ "rewards/GDino": 0.7615998685359955,
1568
+ "rewards/GIT": 0.3799494504928589,
1569
+ "rewards/HPSv2": 0.2450580596923828,
1570
+ "rewards/ORM": 0.36062875390052795,
1571
+ "self_certainty_semantic": -25.5,
1572
+ "self_certainty_token": -21.0625,
1573
+ "step": 92
1574
+ },
1575
+ {
1576
+ "completion_length": 52.5625,
1577
+ "epoch": 0.10299003322259136,
1578
+ "grad_norm": 0.476144403219223,
1579
+ "kl": 0.004547119140625,
1580
+ "learning_rate": 9.41875e-07,
1581
+ "loss": -0.006627652794122696,
1582
+ "reward": 2.3529324531555176,
1583
+ "reward_std": 0.38789400458335876,
1584
+ "rewards/GDino": 0.8122400343418121,
1585
+ "rewards/GIT": 0.40920257568359375,
1586
+ "rewards/HPSv2": 0.25894737243652344,
1587
+ "rewards/ORM": 0.8725424408912659,
1588
+ "self_certainty_semantic": -25.1875,
1589
+ "self_certainty_token": -20.75,
1590
+ "step": 93
1591
+ },
1592
+ {
1593
+ "completion_length": 45.234375,
1594
+ "epoch": 0.10409745293466224,
1595
+ "grad_norm": 0.4303518235683441,
1596
+ "kl": 0.00390625,
1597
+ "learning_rate": 9.4125e-07,
1598
+ "loss": 0.002329372800886631,
1599
+ "reward": 2.063507556915283,
1600
+ "reward_std": 0.4875355362892151,
1601
+ "rewards/GDino": 0.8157378733158112,
1602
+ "rewards/GIT": 0.2162991166114807,
1603
+ "rewards/HPSv2": 0.2860240936279297,
1604
+ "rewards/ORM": 0.7454463839530945,
1605
+ "self_certainty_semantic": -25.3125,
1606
+ "self_certainty_token": -22.25,
1607
+ "step": 94
1608
+ },
1609
+ {
1610
+ "completion_length": 54.21875,
1611
+ "epoch": 0.10520487264673312,
1612
+ "grad_norm": 0.9745371341705322,
1613
+ "kl": 0.004852294921875,
1614
+ "learning_rate": 9.40625e-07,
1615
+ "loss": 0.015892890747636557,
1616
+ "reward": 2.4900766611099243,
1617
+ "reward_std": 0.33158986270427704,
1618
+ "rewards/GDino": 0.9456690549850464,
1619
+ "rewards/GIT": 0.7110534906387329,
1620
+ "rewards/HPSv2": 0.2568836212158203,
1621
+ "rewards/ORM": 0.5764705836772919,
1622
+ "self_certainty_semantic": -25.625,
1623
+ "self_certainty_token": -21.3125,
1624
+ "step": 95
1625
+ },
1626
+ {
1627
+ "completion_length": 62.953125,
1628
+ "epoch": 0.10631229235880399,
1629
+ "grad_norm": 1.6108874082565308,
1630
+ "kl": 0.00475311279296875,
1631
+ "learning_rate": 9.399999999999999e-07,
1632
+ "loss": 0.012537557166069746,
1633
+ "reward": 2.4274561405181885,
1634
+ "reward_std": 0.3028244078159332,
1635
+ "rewards/GDino": 0.9155160486698151,
1636
+ "rewards/GIT": 0.6933247745037079,
1637
+ "rewards/HPSv2": 0.25919437408447266,
1638
+ "rewards/ORM": 0.5594209432601929,
1639
+ "self_certainty_semantic": -25.375,
1640
+ "self_certainty_token": -20.3125,
1641
+ "step": 96
1642
+ },
1643
+ {
1644
+ "completion_length": 44.890625,
1645
+ "epoch": 0.10741971207087486,
1646
+ "grad_norm": 0.42777886986732483,
1647
+ "kl": 0.0064697265625,
1648
+ "learning_rate": 9.393749999999999e-07,
1649
+ "loss": 0.006582918576896191,
1650
+ "reward": 1.7229499220848083,
1651
+ "reward_std": 0.29571742564439774,
1652
+ "rewards/GDino": 0.6976552903652191,
1653
+ "rewards/GIT": 0.17514611035585403,
1654
+ "rewards/HPSv2": 0.2757740020751953,
1655
+ "rewards/ORM": 0.5743745565414429,
1656
+ "self_certainty_semantic": -25.25,
1657
+ "self_certainty_token": -21.5,
1658
+ "step": 97
1659
+ },
1660
+ {
1661
+ "completion_length": 52.78125,
1662
+ "epoch": 0.10852713178294573,
1663
+ "grad_norm": 0.4346785247325897,
1664
+ "kl": 0.00446319580078125,
1665
+ "learning_rate": 9.387499999999999e-07,
1666
+ "loss": 0.010664775501936674,
1667
+ "reward": 1.9896260499954224,
1668
+ "reward_std": 0.5384568274021149,
1669
+ "rewards/GDino": 0.7534899115562439,
1670
+ "rewards/GIT": 0.416723370552063,
1671
+ "rewards/HPSv2": 0.25490760803222656,
1672
+ "rewards/ORM": 0.5645051300525665,
1673
+ "self_certainty_semantic": -25.25,
1674
+ "self_certainty_token": -21.4375,
1675
+ "step": 98
1676
+ },
1677
+ {
1678
+ "completion_length": 53.546875,
1679
+ "epoch": 0.10963455149501661,
1680
+ "grad_norm": 0.4226502478122711,
1681
+ "kl": 0.00543212890625,
1682
+ "learning_rate": 9.381249999999999e-07,
1683
+ "loss": -0.009754271944984794,
1684
+ "reward": 2.1711018085479736,
1685
+ "reward_std": 0.3036491945385933,
1686
+ "rewards/GDino": 0.8239583373069763,
1687
+ "rewards/GIT": 0.6844146698713303,
1688
+ "rewards/HPSv2": 0.24951934814453125,
1689
+ "rewards/ORM": 0.41320937871932983,
1690
+ "self_certainty_semantic": -25.25,
1691
+ "self_certainty_token": -22.0,
1692
+ "step": 99
1693
+ },
1694
+ {
1695
+ "completion_length": 48.0625,
1696
+ "epoch": 0.11074197120708748,
1697
+ "grad_norm": 0.4250389039516449,
1698
+ "kl": 0.00537109375,
1699
+ "learning_rate": 9.374999999999999e-07,
1700
+ "loss": -0.014408082235604525,
1701
+ "reward": 1.9375371932983398,
1702
+ "reward_std": 0.4484590142965317,
1703
+ "rewards/GDino": 0.6897697150707245,
1704
+ "rewards/GIT": 0.4094943553209305,
1705
+ "rewards/HPSv2": 0.2472515106201172,
1706
+ "rewards/ORM": 0.5910216569900513,
1707
+ "self_certainty_semantic": -25.25,
1708
+ "self_certainty_token": -21.4375,
1709
+ "step": 100
1710
+ }
1711
+ ],
1712
+ "logging_steps": 1.0,
1713
+ "max_steps": 1600,
1714
+ "num_input_tokens_seen": 0,
1715
+ "num_train_epochs": 2,
1716
+ "save_steps": 100,
1717
+ "stateful_callbacks": {
1718
+ "TrainerControl": {
1719
+ "args": {
1720
+ "should_epoch_stop": false,
1721
+ "should_evaluate": false,
1722
+ "should_log": false,
1723
+ "should_save": true,
1724
+ "should_training_stop": false
1725
+ },
1726
+ "attributes": {}
1727
+ }
1728
+ },
1729
+ "total_flos": 0.0,
1730
+ "train_batch_size": 1,
1731
+ "trial_name": null,
1732
+ "trial_params": null
1733
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e36f4e9e5eace1020321abae0891c9345612c7ffcda4e80043137b03909411e
3
+ size 10040
zero_to_fp32.py ADDED
@@ -0,0 +1,674 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+
3
+ # Copyright (c) Microsoft Corporation.
4
+ # SPDX-License-Identifier: Apache-2.0
5
+
6
+ # DeepSpeed Team
7
+
8
+ # This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
9
+ # copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
10
+ # the future. Once extracted, the weights don't require DeepSpeed and can be used in any
11
+ # application.
12
+ #
13
+ # example:
14
+ # python zero_to_fp32.py . output_dir/
15
+ # or
16
+ # python zero_to_fp32.py . output_dir/ --safe_serialization
17
+
18
+ import argparse
19
+ import torch
20
+ import glob
21
+ import math
22
+ import os
23
+ import re
24
+ import json
25
+ from tqdm import tqdm
26
+ from collections import OrderedDict
27
+ from dataclasses import dataclass
28
+
29
+ # while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
30
+ # DeepSpeed data structures it has to be available in the current python environment.
31
+ from deepspeed.utils import logger
32
+ from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
33
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
34
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
35
+
36
+
37
+ @dataclass
38
+ class zero_model_state:
39
+ buffers: dict()
40
+ param_shapes: dict()
41
+ shared_params: list
42
+ ds_version: int
43
+ frozen_param_shapes: dict()
44
+ frozen_param_fragments: dict()
45
+
46
+
47
+ debug = 0
48
+
49
+ # load to cpu
50
+ device = torch.device('cpu')
51
+
52
+
53
+ def atoi(text):
54
+ return int(text) if text.isdigit() else text
55
+
56
+
57
+ def natural_keys(text):
58
+ '''
59
+ alist.sort(key=natural_keys) sorts in human order
60
+ http://nedbatchelder.com/blog/200712/human_sorting.html
61
+ (See Toothy's implementation in the comments)
62
+ '''
63
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
64
+
65
+
66
+ def get_model_state_file(checkpoint_dir, zero_stage):
67
+ if not os.path.isdir(checkpoint_dir):
68
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
69
+
70
+ # there should be only one file
71
+ if zero_stage <= 2:
72
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
73
+ elif zero_stage == 3:
74
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
75
+
76
+ if not os.path.exists(file):
77
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
78
+
79
+ return file
80
+
81
+
82
+ def get_checkpoint_files(checkpoint_dir, glob_pattern):
83
+ # XXX: need to test that this simple glob rule works for multi-node setup too
84
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
85
+
86
+ if len(ckpt_files) == 0:
87
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
88
+
89
+ return ckpt_files
90
+
91
+
92
+ def get_optim_files(checkpoint_dir):
93
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
94
+
95
+
96
+ def get_model_state_files(checkpoint_dir):
97
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
98
+
99
+
100
+ def parse_model_states(files):
101
+ zero_model_states = []
102
+ for file in files:
103
+ state_dict = torch.load(file, map_location=device)
104
+
105
+ if BUFFER_NAMES not in state_dict:
106
+ raise ValueError(f"{file} is not a model state checkpoint")
107
+ buffer_names = state_dict[BUFFER_NAMES]
108
+ if debug:
109
+ print("Found buffers:", buffer_names)
110
+
111
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
112
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
113
+ param_shapes = state_dict[PARAM_SHAPES]
114
+
115
+ # collect parameters that are included in param_shapes
116
+ param_names = []
117
+ for s in param_shapes:
118
+ for name in s.keys():
119
+ param_names.append(name)
120
+
121
+ # update with frozen parameters
122
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
123
+ if frozen_param_shapes is not None:
124
+ if debug:
125
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
126
+ param_names += list(frozen_param_shapes.keys())
127
+
128
+ # handle shared params
129
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
130
+
131
+ ds_version = state_dict.get(DS_VERSION, None)
132
+
133
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
134
+
135
+ z_model_state = zero_model_state(buffers=buffers,
136
+ param_shapes=param_shapes,
137
+ shared_params=shared_params,
138
+ ds_version=ds_version,
139
+ frozen_param_shapes=frozen_param_shapes,
140
+ frozen_param_fragments=frozen_param_fragments)
141
+ zero_model_states.append(z_model_state)
142
+
143
+ return zero_model_states
144
+
145
+
146
+ def parse_optim_states(files, ds_checkpoint_dir):
147
+ total_files = len(files)
148
+ state_dicts = []
149
+ for f in files:
150
+ state_dict = torch.load(f, map_location=device)
151
+ # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
152
+ # and also handle the case where it was already removed by another helper script
153
+ state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
154
+ state_dicts.append(state_dict)
155
+
156
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
157
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
158
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
159
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
160
+
161
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
162
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
163
+ # use the max of the partition_count to get the dp world_size.
164
+
165
+ if type(world_size) is list:
166
+ world_size = max(world_size)
167
+
168
+ if world_size != total_files:
169
+ raise ValueError(
170
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
171
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
172
+ )
173
+
174
+ # the groups are named differently in each stage
175
+ if zero_stage <= 2:
176
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
177
+ elif zero_stage == 3:
178
+ fp32_groups_key = FP32_FLAT_GROUPS
179
+ else:
180
+ raise ValueError(f"unknown zero stage {zero_stage}")
181
+
182
+ if zero_stage <= 2:
183
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
184
+ elif zero_stage == 3:
185
+ # if there is more than one param group, there will be multiple flattened tensors - one
186
+ # flattened tensor per group - for simplicity merge them into a single tensor
187
+ #
188
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
189
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
190
+
191
+ fp32_flat_groups = [
192
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
193
+ ]
194
+
195
+ return zero_stage, world_size, fp32_flat_groups
196
+
197
+
198
+ def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
199
+ """
200
+ Returns fp32 state_dict reconstructed from ds checkpoint
201
+
202
+ Args:
203
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
204
+
205
+ """
206
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
207
+
208
+ optim_files = get_optim_files(ds_checkpoint_dir)
209
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
210
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
211
+
212
+ model_files = get_model_state_files(ds_checkpoint_dir)
213
+
214
+ zero_model_states = parse_model_states(model_files)
215
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
216
+
217
+ if zero_stage <= 2:
218
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
219
+ exclude_frozen_parameters)
220
+ elif zero_stage == 3:
221
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
222
+ exclude_frozen_parameters)
223
+
224
+
225
+ def _zero2_merge_frozen_params(state_dict, zero_model_states):
226
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
227
+ return
228
+
229
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
230
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
231
+
232
+ if debug:
233
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
234
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
235
+
236
+ wanted_params = len(frozen_param_shapes)
237
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
238
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
239
+ print(f'Frozen params: Have {avail_numel} numels to process.')
240
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
241
+
242
+ total_params = 0
243
+ total_numel = 0
244
+ for name, shape in frozen_param_shapes.items():
245
+ total_params += 1
246
+ unpartitioned_numel = shape.numel()
247
+ total_numel += unpartitioned_numel
248
+
249
+ state_dict[name] = frozen_param_fragments[name]
250
+
251
+ if debug:
252
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
253
+
254
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
255
+
256
+
257
+ def _has_callable(obj, fn):
258
+ attr = getattr(obj, fn, None)
259
+ return callable(attr)
260
+
261
+
262
+ def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
263
+ param_shapes = zero_model_states[0].param_shapes
264
+
265
+ # Reconstruction protocol:
266
+ #
267
+ # XXX: document this
268
+
269
+ if debug:
270
+ for i in range(world_size):
271
+ for j in range(len(fp32_flat_groups[0])):
272
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
273
+
274
+ # XXX: memory usage doubles here (zero2)
275
+ num_param_groups = len(fp32_flat_groups[0])
276
+ merged_single_partition_of_fp32_groups = []
277
+ for i in range(num_param_groups):
278
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
279
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
280
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
281
+ avail_numel = sum(
282
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
283
+
284
+ if debug:
285
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
286
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
287
+ # not asserting if there is a mismatch due to possible padding
288
+ print(f"Have {avail_numel} numels to process.")
289
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
290
+
291
+ # params
292
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
293
+ # out-of-core computing solution
294
+ total_numel = 0
295
+ total_params = 0
296
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
297
+ offset = 0
298
+ avail_numel = full_single_fp32_vector.numel()
299
+ for name, shape in shapes.items():
300
+
301
+ unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
302
+ total_numel += unpartitioned_numel
303
+ total_params += 1
304
+
305
+ if debug:
306
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
307
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
308
+ offset += unpartitioned_numel
309
+
310
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
311
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
312
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
313
+ # live optimizer object, so we are checking that the numbers are within the right range
314
+ align_to = 2 * world_size
315
+
316
+ def zero2_align(x):
317
+ return align_to * math.ceil(x / align_to)
318
+
319
+ if debug:
320
+ print(f"original offset={offset}, avail_numel={avail_numel}")
321
+
322
+ offset = zero2_align(offset)
323
+ avail_numel = zero2_align(avail_numel)
324
+
325
+ if debug:
326
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
327
+
328
+ # Sanity check
329
+ if offset != avail_numel:
330
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
331
+
332
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
333
+
334
+
335
+ def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
336
+ exclude_frozen_parameters):
337
+ state_dict = OrderedDict()
338
+
339
+ # buffers
340
+ buffers = zero_model_states[0].buffers
341
+ state_dict.update(buffers)
342
+ if debug:
343
+ print(f"added {len(buffers)} buffers")
344
+
345
+ if not exclude_frozen_parameters:
346
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
347
+
348
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
349
+
350
+ # recover shared parameters
351
+ for pair in zero_model_states[0].shared_params:
352
+ if pair[1] in state_dict:
353
+ state_dict[pair[0]] = state_dict[pair[1]]
354
+
355
+ return state_dict
356
+
357
+
358
+ def zero3_partitioned_param_info(unpartitioned_numel, world_size):
359
+ remainder = unpartitioned_numel % world_size
360
+ padding_numel = (world_size - remainder) if remainder else 0
361
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
362
+ return partitioned_numel, padding_numel
363
+
364
+
365
+ def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
366
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
367
+ return
368
+
369
+ if debug:
370
+ for i in range(world_size):
371
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
372
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
373
+
374
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
375
+ wanted_params = len(frozen_param_shapes)
376
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
377
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
378
+ print(f'Frozen params: Have {avail_numel} numels to process.')
379
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
380
+
381
+ total_params = 0
382
+ total_numel = 0
383
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
384
+ total_params += 1
385
+ unpartitioned_numel = shape.numel()
386
+ total_numel += unpartitioned_numel
387
+
388
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
389
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
390
+
391
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
392
+
393
+ if debug:
394
+ print(
395
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
396
+ )
397
+
398
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
399
+
400
+
401
+ def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
402
+ param_shapes = zero_model_states[0].param_shapes
403
+ avail_numel = fp32_flat_groups[0].numel() * world_size
404
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
405
+ # param, re-consolidating each param, while dealing with padding if any
406
+
407
+ # merge list of dicts, preserving order
408
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
409
+
410
+ if debug:
411
+ for i in range(world_size):
412
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
413
+
414
+ wanted_params = len(param_shapes)
415
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
416
+ # not asserting if there is a mismatch due to possible padding
417
+ avail_numel = fp32_flat_groups[0].numel() * world_size
418
+ print(f"Trainable params: Have {avail_numel} numels to process.")
419
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
420
+
421
+ # params
422
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
423
+ # out-of-core computing solution
424
+ offset = 0
425
+ total_numel = 0
426
+ total_params = 0
427
+ for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
428
+ unpartitioned_numel = shape.numel()
429
+ total_numel += unpartitioned_numel
430
+ total_params += 1
431
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
432
+
433
+ if debug:
434
+ print(
435
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
436
+ )
437
+
438
+ # XXX: memory usage doubles here
439
+ state_dict[name] = torch.cat(
440
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
441
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
442
+ offset += partitioned_numel
443
+
444
+ offset *= world_size
445
+
446
+ # Sanity check
447
+ if offset != avail_numel:
448
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
449
+
450
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
451
+
452
+
453
+ def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
454
+ exclude_frozen_parameters):
455
+ state_dict = OrderedDict()
456
+
457
+ # buffers
458
+ buffers = zero_model_states[0].buffers
459
+ state_dict.update(buffers)
460
+ if debug:
461
+ print(f"added {len(buffers)} buffers")
462
+
463
+ if not exclude_frozen_parameters:
464
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
465
+
466
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
467
+
468
+ # recover shared parameters
469
+ for pair in zero_model_states[0].shared_params:
470
+ if pair[1] in state_dict:
471
+ state_dict[pair[0]] = state_dict[pair[1]]
472
+
473
+ return state_dict
474
+
475
+
476
+ def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
477
+ """
478
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
479
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
480
+ via a model hub.
481
+
482
+ Args:
483
+ - ``checkpoint_dir``: path to the desired checkpoint folder
484
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
485
+ - ``exclude_frozen_parameters``: exclude frozen parameters
486
+
487
+ Returns:
488
+ - pytorch ``state_dict``
489
+
490
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
491
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
492
+ the checkpoint.
493
+
494
+ A typical usage might be ::
495
+
496
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
497
+ # do the training and checkpoint saving
498
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
499
+ model = model.cpu() # move to cpu
500
+ model.load_state_dict(state_dict)
501
+ # submit to model hub or save the model to share with others
502
+
503
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
504
+ application. i.e. you will need to re-initialize the deepspeed engine, since
505
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
506
+
507
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
508
+
509
+ """
510
+ if tag is None:
511
+ latest_path = os.path.join(checkpoint_dir, 'latest')
512
+ if os.path.isfile(latest_path):
513
+ with open(latest_path, 'r') as fd:
514
+ tag = fd.read().strip()
515
+ else:
516
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
517
+
518
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
519
+
520
+ if not os.path.isdir(ds_checkpoint_dir):
521
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
522
+
523
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
524
+
525
+
526
+ def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
527
+ output_dir,
528
+ max_shard_size="5GB",
529
+ safe_serialization=False,
530
+ tag=None,
531
+ exclude_frozen_parameters=False):
532
+ """
533
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
534
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
535
+
536
+ Args:
537
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
538
+ - ``output_dir``: directory to the pytorch fp32 state_dict output files
539
+ - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
540
+ - ``safe_serialization``: whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
541
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
542
+ - ``exclude_frozen_parameters``: exclude frozen parameters
543
+ """
544
+ # Dependency pre-check
545
+ if safe_serialization:
546
+ try:
547
+ from safetensors.torch import save_file
548
+ except ImportError:
549
+ print('If you want to use `safe_serialization`, please `pip install safetensors`')
550
+ raise
551
+ if max_shard_size is not None:
552
+ try:
553
+ from huggingface_hub import split_torch_state_dict_into_shards
554
+ except ImportError:
555
+ print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
556
+ raise
557
+
558
+ # Convert zero checkpoint to state_dict
559
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
560
+
561
+ # Shard the model if it is too big.
562
+ weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
563
+ if max_shard_size is not None:
564
+ filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
565
+ state_dict_split = split_torch_state_dict_into_shards(state_dict,
566
+ filename_pattern=filename_pattern,
567
+ max_shard_size=max_shard_size)
568
+ else:
569
+ from collections import namedtuple
570
+ StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
571
+ state_dict_split = StateDictSplit(is_sharded=False,
572
+ filename_to_tensors={weights_name: list(state_dict.keys())})
573
+
574
+ # Save the model
575
+ filename_to_tensors = state_dict_split.filename_to_tensors.items()
576
+ for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
577
+ shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
578
+ output_path = os.path.join(output_dir, shard_file)
579
+ if safe_serialization:
580
+ save_file(shard, output_path, metadata={"format": "pt"})
581
+ else:
582
+ torch.save(shard, output_path)
583
+
584
+ # Save index if sharded
585
+ if state_dict_split.is_sharded:
586
+ index = {
587
+ "metadata": state_dict_split.metadata,
588
+ "weight_map": state_dict_split.tensor_to_filename,
589
+ }
590
+ save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
591
+ save_index_file = os.path.join(output_dir, save_index_file)
592
+ with open(save_index_file, "w", encoding="utf-8") as f:
593
+ content = json.dumps(index, indent=2, sort_keys=True) + "\n"
594
+ f.write(content)
595
+
596
+
597
+ def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
598
+ """
599
+ 1. Put the provided model to cpu
600
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
601
+ 3. Load it into the provided model
602
+
603
+ Args:
604
+ - ``model``: the model object to update
605
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
606
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
607
+
608
+ Returns:
609
+ - ``model`: modified model
610
+
611
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
612
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
613
+ conveniently placed for you in the checkpoint folder.
614
+
615
+ A typical usage might be ::
616
+
617
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
618
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
619
+ # submit to model hub or save the model to share with others
620
+
621
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
622
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
623
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
624
+
625
+ """
626
+ logger.info(f"Extracting fp32 weights")
627
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
628
+
629
+ logger.info(f"Overwriting model with fp32 weights")
630
+ model = model.cpu()
631
+ model.load_state_dict(state_dict, strict=False)
632
+
633
+ return model
634
+
635
+
636
+ if __name__ == "__main__":
637
+ parser = argparse.ArgumentParser()
638
+ parser.add_argument("checkpoint_dir",
639
+ type=str,
640
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
641
+ parser.add_argument("output_dir",
642
+ type=str,
643
+ help="directory to the pytorch fp32 state_dict output files"
644
+ "(e.g. path/checkpoint-12-output/)")
645
+ parser.add_argument(
646
+ "--max_shard_size",
647
+ type=str,
648
+ default="5GB",
649
+ help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
650
+ "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
651
+ "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
652
+ "without CPU OOM issues.")
653
+ parser.add_argument(
654
+ "--safe_serialization",
655
+ default=False,
656
+ action='store_true',
657
+ help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
658
+ parser.add_argument("-t",
659
+ "--tag",
660
+ type=str,
661
+ default=None,
662
+ help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
663
+ parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
664
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
665
+ args = parser.parse_args()
666
+
667
+ debug = args.debug
668
+
669
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
670
+ args.output_dir,
671
+ max_shard_size=args.max_shard_size,
672
+ safe_serialization=args.safe_serialization,
673
+ tag=args.tag,
674
+ exclude_frozen_parameters=args.exclude_frozen_parameters)