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
1f298b5
·
verified ·
1 Parent(s): d024fdc

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-108000/config.json +79 -0
  3. checkpoint-108000/generation_config.json +6 -0
  4. checkpoint-108000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-108000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-108000/model.safetensors.index.json +0 -0
  7. checkpoint-108000/optimizer.pt +3 -0
  8. checkpoint-108000/rng_state_0.pth +3 -0
  9. checkpoint-108000/rng_state_1.pth +3 -0
  10. checkpoint-108000/rng_state_2.pth +3 -0
  11. checkpoint-108000/rng_state_3.pth +3 -0
  12. checkpoint-108000/scheduler.pt +3 -0
  13. checkpoint-108000/trainer_state.json +3 -0
  14. checkpoint-108000/training_args.bin +3 -0
  15. checkpoint-110000/config.json +79 -0
  16. checkpoint-110000/generation_config.json +6 -0
  17. checkpoint-110000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-110000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-110000/model.safetensors.index.json +0 -0
  20. checkpoint-110000/optimizer.pt +3 -0
  21. checkpoint-110000/rng_state_0.pth +3 -0
  22. checkpoint-110000/rng_state_1.pth +3 -0
  23. checkpoint-110000/rng_state_2.pth +3 -0
  24. checkpoint-110000/rng_state_3.pth +3 -0
  25. checkpoint-110000/scheduler.pt +3 -0
  26. checkpoint-110000/trainer_state.json +3 -0
  27. checkpoint-110000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -40,3 +40,5 @@ checkpoint-88000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
40
  checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
41
  checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
42
  checkpoint-98000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
40
  checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
41
  checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
42
  checkpoint-98000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
43
+ checkpoint-108000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
44
+ checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-108000/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-108000/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-108000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2cdf0bcddad4eef696d82820d924f65a895414ece91da756baacffaf2c5f52c
3
+ size 4974360904
checkpoint-108000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:756a25daa806839e1018a7bb1e7d991966f34c02baeb93c1ce51f85458105855
3
+ size 3234250886
checkpoint-108000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-108000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eec38185af072a1f1f1513d872fd59987de59698f2b43dc0d965a69e861c3d6f
3
+ size 7455061885
checkpoint-108000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f4e4d04b52c7e255ea3f4f8b2813e412eb12dc8b3a3c7730a6d78923c326303
3
+ size 15024
checkpoint-108000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a844a31569a83cbe0ef181dad5d01a4b9c4836b64d2b5ce9ec6c48117fbb4664
3
+ size 15024
checkpoint-108000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:167ef72385220ad82ecf7be1802f370a9615d1c68c973096e9cb8109b728689e
3
+ size 15024
checkpoint-108000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dabf6b2c5be108f56a58bd9967ef9f88e63f28e473c482d0eca73b4222fc14f5
3
+ size 15024
checkpoint-108000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a73d692633fae83ed5d4ea37065103cf8fe9b70abb3c760b30a55df16438f43
3
+ size 1064
checkpoint-108000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4e2f122919b8a0128f7580e8a94250c4da18b25044a326a67106086fe337a41
3
+ size 16611236
checkpoint-108000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:234021f9ce0ba25ff91165f39124617fc0a90a6a0f8f32fe1ce4a525be4dcb88
3
+ size 6008
checkpoint-110000/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-110000/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-110000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c3ae8ad74a7bdbb6bcab929cba165fda0b8dd250ed412a708646250fa9e10e9
3
+ size 4974360904
checkpoint-110000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:800d3b43534b57681f5eb797f7b2464356165308eb307ab88be60be8e53faf1f
3
+ size 3234250886
checkpoint-110000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-110000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd9823a067e1730d2879627d77b610f5af0d721636dedb0790ab1de173b4d3cc
3
+ size 7455061885
checkpoint-110000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d895f1bb5b960f0ddb1bf259634b287bd395f50ffc7c9f4782f89ade47e3a76
3
+ size 15024
checkpoint-110000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f502c7eae13d4a130dafaf8e011e5794bc47ca2f8a2ce85e7d6e0215d66a5e9
3
+ size 15024
checkpoint-110000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:460f84de9a4f888cd5e989e5ad5ecb10fe6d921fa661f20c381ba9de1b8c6f48
3
+ size 15024
checkpoint-110000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:570c9db855774b5b07c28661e9a66ed44fe96ab5860f107c63ca27f62f829874
3
+ size 15024
checkpoint-110000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c4f0b2439dd7fa4ae2aceb98ade9f541d95d2f057f7ce4f443e7cd4efa5cd45
3
+ size 1064
checkpoint-110000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf778a97c254171c5783c9b20453daf380f6f46df655fe816389d8d5e512b47c
3
+ size 16920019
checkpoint-110000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:234021f9ce0ba25ff91165f39124617fc0a90a6a0f8f32fe1ce4a525be4dcb88
3
+ size 6008