Hugging Face's logo

zhiyang1
/
blip3o_f32c32-sana-1.0-768_patch-1_group-7_fusion_residual_attn

Safetensors
Model card Files
xet
Community
zhiyang1 commited on
Commit
29466c8
·
verified ·
1 Parent(s): 8926b29

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-268000/config.json +79 -0
  3. checkpoint-268000/generation_config.json +6 -0
  4. checkpoint-268000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-268000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-268000/model.safetensors.index.json +0 -0
  7. checkpoint-268000/optimizer.pt +3 -0
  8. checkpoint-268000/rng_state_0.pth +3 -0
  9. checkpoint-268000/rng_state_1.pth +3 -0
  10. checkpoint-268000/rng_state_2.pth +3 -0
  11. checkpoint-268000/rng_state_3.pth +3 -0
  12. checkpoint-268000/scheduler.pt +3 -0
  13. checkpoint-268000/trainer_state.json +3 -0
  14. checkpoint-268000/training_args.bin +3 -0
  15. checkpoint-270000/config.json +79 -0
  16. checkpoint-270000/generation_config.json +6 -0
  17. checkpoint-270000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-270000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-270000/model.safetensors.index.json +0 -0
  20. checkpoint-270000/optimizer.pt +3 -0
  21. checkpoint-270000/rng_state_0.pth +3 -0
  22. checkpoint-270000/rng_state_1.pth +3 -0
  23. checkpoint-270000/rng_state_2.pth +3 -0
  24. checkpoint-270000/rng_state_3.pth +3 -0
  25. checkpoint-270000/scheduler.pt +3 -0
  26. checkpoint-270000/trainer_state.json +3 -0
  27. checkpoint-270000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -113,3 +113,5 @@ checkpoint-248000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
113
  checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
114
  checkpoint-258000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
115
  checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
113
  checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
114
  checkpoint-258000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
115
  checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
116
+ checkpoint-268000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
117
+ checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-268000/config.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "ar_steps": 1,
3
+ "architectures": [
4
+ "DiffVLMDiffusion"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151643,
8
+ "condition_layer": -1,
9
+ "eos_token_id": 151645,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 1536,
12
+ "image_token_id": 151655,
13
+ "img_cross_attention_dim": 2048,
14
+ "img_diffuser_depth": 1,
15
+ "img_ffn_dim_multiplier": null,
16
+ "img_hidden_size": 1536,
17
+ "img_multiple_of": 256,
18
+ "img_norm_eps": 1e-05,
19
+ "img_num_attention_heads": 12,
20
+ "img_num_kv_heads": 12,
21
+ "img_qk_norm": true,
22
+ "in_channels": 32,
23
+ "initializer_range": 0.02,
24
+ "inject_img_diffuser": false,
25
+ "input_size": 32,
26
+ "intermediate_size": 8960,
27
+ "layer_group_size": 7,
28
+ "layerwise_start_idx": 0,
29
+ "lora_alpha": 16,
30
+ "lora_bias": "none",
31
+ "lora_dropout": 0.05,
32
+ "lora_enable": false,
33
+ "lora_r": 64,
34
+ "max_position_embeddings": 32768,
35
+ "max_window_layers": 28,
36
+ "model_type": "qwen2_vl",
37
+ "non_linearity": 1,
38
+ "norm_elementwise_affine": true,
39
+ "num_attention_heads": 12,
40
+ "num_hidden_layers": 28,
41
+ "num_key_value_heads": 2,
42
+ "patch_size": 1,
43
+ "repa_coeff": 0.5,
44
+ "repa_layers": null,
45
+ "repa_shared": false,
46
+ "rms_norm_eps": 1e-06,
47
+ "rope_scaling": {
48
+ "mrope_section": [
49
+ 16,
50
+ 24,
51
+ 24
52
+ ],
53
+ "rope_type": "default",
54
+ "type": "default"
55
+ },
56
+ "rope_theta": 1000000.0,
57
+ "sample_size": 128,
58
+ "sampling_steps": 28,
59
+ "sliding_window": null,
60
+ "tie_word_embeddings": true,
61
+ "torch_dtype": "bfloat16",
62
+ "transformers_version": "4.47.0",
63
+ "use_cache": true,
64
+ "use_repa": false,
65
+ "use_residual_attn": true,
66
+ "use_sliding_window": false,
67
+ "vae_path": "mit-han-lab/dc-ae-f32c32-sana-1.1-diffusers",
68
+ "video_token_id": 151656,
69
+ "vision_config": {
70
+ "hidden_size": 1536,
71
+ "in_chans": 3,
72
+ "model_type": "qwen2_vl",
73
+ "spatial_patch_size": 14
74
+ },
75
+ "vision_end_token_id": 151653,
76
+ "vision_start_token_id": 151652,
77
+ "vision_token_id": 151654,
78
+ "vocab_size": 151936
79
+ }
checkpoint-268000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 151643,
4
+ "eos_token_id": 151645,
5
+ "transformers_version": "4.47.0"
6
+ }
checkpoint-268000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7856be896c72f3961ce2442e311de918763e4854a8c052bea831bc0d2678f07
3
+ size 4974360904
checkpoint-268000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2867c0d7aca11447847e26211c1127b87fa38e2c20e7496891ef01237ba8cf2b
3
+ size 3234250886
checkpoint-268000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-268000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:686349378bc6557fa3579148503c0799b5e405dc4b31999de4b47616f3027e96
3
+ size 7455061885
checkpoint-268000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6df3233357fd9e1d17cfc1afba3b7787f844705bc94958db36f570b9359afbf
3
+ size 15024
checkpoint-268000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ddfcc54e0f8ff954b8817a219ac7f1e6b6a964f7573b8a82590d4c3486ee31d
3
+ size 15024
checkpoint-268000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dabdf4d1c2537b4612f970dfd78676a7cd7ea88cf8f2966d0ad80cc2ae4d3aac
3
+ size 15024
checkpoint-268000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bec18cff682451f7757154ba20ec828b478fb72e850ebc1dd002dd2e53602be1
3
+ size 15024
checkpoint-268000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:202cf6f6e3ab774826cf6e94911af1995881a049a72cde169663071c0eb2d77a
3
+ size 1064
checkpoint-268000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09632cd366f848cbff63fef406ce9fc8795d5f7f3479032dd9972912987544cb
3
+ size 41303790
checkpoint-268000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c50b27b5df5190f876cc484d1f8e08d53cce23b2b7fb94669441e8292e002ed8
3
+ size 6072
checkpoint-270000/config.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "ar_steps": 1,
3
+ "architectures": [
4
+ "DiffVLMDiffusion"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151643,
8
+ "condition_layer": -1,
9
+ "eos_token_id": 151645,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 1536,
12
+ "image_token_id": 151655,
13
+ "img_cross_attention_dim": 2048,
14
+ "img_diffuser_depth": 1,
15
+ "img_ffn_dim_multiplier": null,
16
+ "img_hidden_size": 1536,
17
+ "img_multiple_of": 256,
18
+ "img_norm_eps": 1e-05,
19
+ "img_num_attention_heads": 12,
20
+ "img_num_kv_heads": 12,
21
+ "img_qk_norm": true,
22
+ "in_channels": 32,
23
+ "initializer_range": 0.02,
24
+ "inject_img_diffuser": false,
25
+ "input_size": 32,
26
+ "intermediate_size": 8960,
27
+ "layer_group_size": 7,
28
+ "layerwise_start_idx": 0,
29
+ "lora_alpha": 16,
30
+ "lora_bias": "none",
31
+ "lora_dropout": 0.05,
32
+ "lora_enable": false,
33
+ "lora_r": 64,
34
+ "max_position_embeddings": 32768,
35
+ "max_window_layers": 28,
36
+ "model_type": "qwen2_vl",
37
+ "non_linearity": 1,
38
+ "norm_elementwise_affine": true,
39
+ "num_attention_heads": 12,
40
+ "num_hidden_layers": 28,
41
+ "num_key_value_heads": 2,
42
+ "patch_size": 1,
43
+ "repa_coeff": 0.5,
44
+ "repa_layers": null,
45
+ "repa_shared": false,
46
+ "rms_norm_eps": 1e-06,
47
+ "rope_scaling": {
48
+ "mrope_section": [
49
+ 16,
50
+ 24,
51
+ 24
52
+ ],
53
+ "rope_type": "default",
54
+ "type": "default"
55
+ },
56
+ "rope_theta": 1000000.0,
57
+ "sample_size": 128,
58
+ "sampling_steps": 28,
59
+ "sliding_window": null,
60
+ "tie_word_embeddings": true,
61
+ "torch_dtype": "bfloat16",
62
+ "transformers_version": "4.47.0",
63
+ "use_cache": true,
64
+ "use_repa": false,
65
+ "use_residual_attn": true,
66
+ "use_sliding_window": false,
67
+ "vae_path": "mit-han-lab/dc-ae-f32c32-sana-1.1-diffusers",
68
+ "video_token_id": 151656,
69
+ "vision_config": {
70
+ "hidden_size": 1536,
71
+ "in_chans": 3,
72
+ "model_type": "qwen2_vl",
73
+ "spatial_patch_size": 14
74
+ },
75
+ "vision_end_token_id": 151653,
76
+ "vision_start_token_id": 151652,
77
+ "vision_token_id": 151654,
78
+ "vocab_size": 151936
79
+ }
checkpoint-270000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 151643,
4
+ "eos_token_id": 151645,
5
+ "transformers_version": "4.47.0"
6
+ }
checkpoint-270000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71a11d68a297d6221466c85a223d218bff67586ffade33fdd928a6d9e7c9b71d
3
+ size 4974360904
checkpoint-270000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c54e149a3085c0025c6761000bab1fd3f43563fd9835e170af6ee3774a333ce
3
+ size 3234250886
checkpoint-270000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-270000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f93fe171f5edc229099e4cfe51c2882377c91c5b511008414b4f54cbee1a653
3
+ size 7455061885
checkpoint-270000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cde7e3ecbf1e310c917b3b39086fc36550ba20352f65f1e70d8b0ac388da703
3
+ size 15024
checkpoint-270000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fe007fd8c9007f4c770f9db55dbfcab23364a0ffbe14b8e40927a5e2e01c0df
3
+ size 15024
checkpoint-270000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef4ff70d39816cbcf34b07dbcaf34e892d65e108cbaa7057be7d065feff9a9d1
3
+ size 15024
checkpoint-270000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e3fab9b140586137747003ccd72fc7f09dc9939c7ae2bd42dbbf9de9399be05
3
+ size 15024
checkpoint-270000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6280f55eef03ab965fa60ac96072d974f287cb3e5f2146b72fed3cb30c22a45
3
+ size 1064
checkpoint-270000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa0c9ca2b59fa7dfd7c55e89166f15e4de91aa8022de1a28a23e1b37317837bd
3
+ size 41612063
checkpoint-270000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c50b27b5df5190f876cc484d1f8e08d53cce23b2b7fb94669441e8292e002ed8
3
+ size 6072