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
d024fdc
·
verified ·
1 Parent(s): 6470fc2

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-100000/config.json +79 -0
  3. checkpoint-100000/generation_config.json +6 -0
  4. checkpoint-100000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-100000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-100000/model.safetensors.index.json +0 -0
  7. checkpoint-100000/optimizer.pt +3 -0
  8. checkpoint-100000/rng_state_0.pth +3 -0
  9. checkpoint-100000/rng_state_1.pth +3 -0
  10. checkpoint-100000/rng_state_2.pth +3 -0
  11. checkpoint-100000/rng_state_3.pth +3 -0
  12. checkpoint-100000/scheduler.pt +3 -0
  13. checkpoint-100000/trainer_state.json +3 -0
  14. checkpoint-100000/training_args.bin +3 -0
  15. checkpoint-98000/config.json +79 -0
  16. checkpoint-98000/generation_config.json +6 -0
  17. checkpoint-98000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-98000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-98000/model.safetensors.index.json +0 -0
  20. checkpoint-98000/optimizer.pt +3 -0
  21. checkpoint-98000/rng_state_0.pth +3 -0
  22. checkpoint-98000/rng_state_1.pth +3 -0
  23. checkpoint-98000/rng_state_2.pth +3 -0
  24. checkpoint-98000/rng_state_3.pth +3 -0
  25. checkpoint-98000/scheduler.pt +3 -0
  26. checkpoint-98000/trainer_state.json +3 -0
  27. checkpoint-98000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -38,3 +38,5 @@ checkpoint-78000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
38
  checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
39
  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
 
 
 
38
  checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
39
  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
checkpoint-100000/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-100000/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-100000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78e84b3ba9dbb6d064568635a8bbd673c18c7d73012acdb0e1e9193d2d9e24b8
3
+ size 4974360904
checkpoint-100000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:963aa57b7788ed5f40ed223e8a8db96c4747c22cdc927c0fc47d888b66b94fea
3
+ size 3234250886
checkpoint-100000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-100000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cede0e05588716763fce390b653d9ea4043f6da8d8baac7a4e605956cc610be
3
+ size 7455061885
checkpoint-100000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3dbc55b4e27781d1ae4805a33f8fcc9f1fff969151e0b65289ffef4912742fb6
3
+ size 15024
checkpoint-100000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:690b031db7736783975b6c51bd3d1fa7dc3b88bf40ae7e149f2eae81257f7610
3
+ size 15024
checkpoint-100000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66c36dc60dadaad1f88403f4b03c47e8acb43a480a1c4f346309375a9c35a0e3
3
+ size 15024
checkpoint-100000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fccb1c65fc1cdc4d5ac88680388742bc370dc74dcf7c9d634181ad8f8fa4e882
3
+ size 15024
checkpoint-100000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:596369a6f0297594b75a7f1b0e6307c25aad34ba553687e462226a0d47547959
3
+ size 1064
checkpoint-100000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9184044973242f117c84a8469fd3e836e3a98b7134c387505f33766697d2c7bf
3
+ size 15375089
checkpoint-100000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:542d7dfa59207a554f97f07971101319d52c8c9c403454744eec74f4a5492ded
3
+ size 6008
checkpoint-98000/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-98000/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-98000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6e8a461126d8f67fd9cd7bb70b8ad9503a7943409c6e2beeaf842f90ea98ff5
3
+ size 4974360904
checkpoint-98000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ce9f2ca6ecea1d3b0c76fa0aa6e3805e04da9bfa0970908d5ebd888418fbec1
3
+ size 3234250886
checkpoint-98000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-98000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6461cc4b39b8e566947784ec6930cd09d81c6d5c91db7b577bbfd2e89c347f24
3
+ size 7455061885
checkpoint-98000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d50b71f620bade28e773525a3c3e92d9f180d11d4f1b39bbab668466dda3a0d6
3
+ size 15024
checkpoint-98000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2b890f75b18f391373a2ad8bd8faa900e57db4bf3f1fd7e7ff518db47f9b993
3
+ size 15024
checkpoint-98000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0eeaa894a11840bc2f6ceff73b57b4b41708a94aea47295207299ef1eb5fdf5b
3
+ size 15024
checkpoint-98000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:407b6baa22cc678f96c09bb9a7593386d50187d37b6ad08daf4721bb17e5786d
3
+ size 15024
checkpoint-98000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34b681a55ba6ca82132ef6045dd5311f7ef9ed33df0a7d214b085e2ceadc690e
3
+ size 1064
checkpoint-98000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a58768aaccad1896997049e2467e0ce2899e297ccc7cb9a0ba0179b4d67271b7
3
+ size 15067990
checkpoint-98000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:542d7dfa59207a554f97f07971101319d52c8c9c403454744eec74f4a5492ded
3
+ size 6008