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
91a73c2
·
verified ·
1 Parent(s): 9fe1a97

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-128000/config.json +79 -0
  3. checkpoint-128000/generation_config.json +6 -0
  4. checkpoint-128000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-128000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-128000/model.safetensors.index.json +0 -0
  7. checkpoint-128000/optimizer.pt +3 -0
  8. checkpoint-128000/rng_state_0.pth +3 -0
  9. checkpoint-128000/rng_state_1.pth +3 -0
  10. checkpoint-128000/rng_state_2.pth +3 -0
  11. checkpoint-128000/rng_state_3.pth +3 -0
  12. checkpoint-128000/scheduler.pt +3 -0
  13. checkpoint-128000/trainer_state.json +3 -0
  14. checkpoint-128000/training_args.bin +3 -0
  15. checkpoint-130000/config.json +79 -0
  16. checkpoint-130000/generation_config.json +6 -0
  17. checkpoint-130000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-130000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-130000/model.safetensors.index.json +0 -0
  20. checkpoint-130000/optimizer.pt +3 -0
  21. checkpoint-130000/rng_state_0.pth +3 -0
  22. checkpoint-130000/rng_state_1.pth +3 -0
  23. checkpoint-130000/rng_state_2.pth +3 -0
  24. checkpoint-130000/rng_state_3.pth +3 -0
  25. checkpoint-130000/scheduler.pt +3 -0
  26. checkpoint-130000/trainer_state.json +3 -0
  27. checkpoint-130000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -44,3 +44,5 @@ 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
45
  checkpoint-118000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
46
  checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
44
  checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
45
  checkpoint-118000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
46
  checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
47
+ checkpoint-128000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
48
+ checkpoint-130000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-128000/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-128000/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-128000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4fceaf8f4d628cf30308f43757aca4ae113c17e5e490ad5adf5a75faed963b77
3
+ size 4974360904
checkpoint-128000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6e581e071f81303c63a95ffe36be5b1fde78eb0b77d64d20019601c18831746
3
+ size 3234250886
checkpoint-128000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-128000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4dea846e1b48320d9ad651e9ec726bad9e172d50ad6e39a2da5dfcfdc91a9098
3
+ size 7455061885
checkpoint-128000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:402867c2a558620633a5703c3f6d04bdb8ebed9b096bf407aa5ac0c6e4948a2b
3
+ size 15024
checkpoint-128000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbbda2613bbd7b944980ab0ea23051d6238b98c8391578f36d1c9525ad1f6012
3
+ size 15024
checkpoint-128000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef0150cf553122d0b8bcea3a609b2f0f1d3695ea7e8943913b4ce1f8f9b8b67c
3
+ size 15024
checkpoint-128000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c9f5c1b41cff950db133f1b03a480a5ac870c1a59988c65dc813387206d0c35
3
+ size 15024
checkpoint-128000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19941965b04b4d5cbb744ba30c5b70491200722b3e649636992da905e05aa033
3
+ size 1064
checkpoint-128000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bc935f90224acb3e41714ac59c77e8ccefb069efae38894a335001acd765984
3
+ size 19699330
checkpoint-128000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5269507d5c0e2a1f3cadf7688c5e43a657393238bc578d0cca887e89090711b3
3
+ size 6008
checkpoint-130000/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-130000/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-130000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bebacb9f40fa5b04745f3519d18f5a425a118400f4d090b76a12262089f3cf5d
3
+ size 4974360904
checkpoint-130000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d4dbd6db14d46b1d1175afd8f7a9950cb6a1b2826a078ea7909e5e36c4ebb42
3
+ size 3234250886
checkpoint-130000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-130000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b4fd3ce6a2b39dc46c6cfd62d7707b4842f3069fbc2a9217563fa6362faaeab
3
+ size 7455061885
checkpoint-130000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51e24e20c807efb00912f2e3bfafd8077b8b0c187ff94813f36dc3e55412095a
3
+ size 15024
checkpoint-130000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3dc29bd9207053e83c9c1d636d5c16d3b64a242ae7d6091ca7b46bb054ddcd4
3
+ size 15024
checkpoint-130000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ca03adbcc6e4c07867543d68219b77ae4624fe4c2173669bb224948483228b7
3
+ size 15024
checkpoint-130000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec43f7a17247a68edf28cc1ebff550320f79126bb7807c373a0b30cef7b2a3d7
3
+ size 15024
checkpoint-130000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ff5845dcb64356c52a4b15f7f2820cd92de83b2220edc96ceea87b3715478bb
3
+ size 1064
checkpoint-130000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4ffa26f82b3e980078d475107b3a0bf76b46750207c97c1bd6b634960da2503
3
+ size 20008169
checkpoint-130000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5269507d5c0e2a1f3cadf7688c5e43a657393238bc578d0cca887e89090711b3
3
+ size 6008