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
6aea61a
·
verified ·
1 Parent(s): 9aa9931

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-158000/config.json +79 -0
  3. checkpoint-158000/generation_config.json +6 -0
  4. checkpoint-158000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-158000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-158000/model.safetensors.index.json +0 -0
  7. checkpoint-158000/optimizer.pt +3 -0
  8. checkpoint-158000/rng_state_0.pth +3 -0
  9. checkpoint-158000/rng_state_1.pth +3 -0
  10. checkpoint-158000/rng_state_2.pth +3 -0
  11. checkpoint-158000/rng_state_3.pth +3 -0
  12. checkpoint-158000/scheduler.pt +3 -0
  13. checkpoint-158000/trainer_state.json +3 -0
  14. checkpoint-158000/training_args.bin +3 -0
  15. checkpoint-160000/config.json +79 -0
  16. checkpoint-160000/generation_config.json +6 -0
  17. checkpoint-160000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-160000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-160000/model.safetensors.index.json +0 -0
  20. checkpoint-160000/optimizer.pt +3 -0
  21. checkpoint-160000/rng_state_0.pth +3 -0
  22. checkpoint-160000/rng_state_1.pth +3 -0
  23. checkpoint-160000/rng_state_2.pth +3 -0
  24. checkpoint-160000/rng_state_3.pth +3 -0
  25. checkpoint-160000/scheduler.pt +3 -0
  26. checkpoint-160000/trainer_state.json +3 -0
  27. checkpoint-160000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -76,3 +76,5 @@ checkpoint-138000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
76
  checkpoint-140000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
77
  checkpoint-148000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
78
  checkpoint-150000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
76
  checkpoint-140000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
77
  checkpoint-148000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
78
  checkpoint-150000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
79
+ checkpoint-158000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
80
+ checkpoint-160000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-158000/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-158000/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-158000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7111245aef0b7934a6f0a45e3b3d197df6eb5ac727665a0ee104e712a06d05a
3
+ size 4974360904
checkpoint-158000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3334dc7f3cce8059aacfc2b5feb7a22bf951636c661688d389a9206d4af3fd2f
3
+ size 3234250886
checkpoint-158000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-158000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a94464c6c542a08b350a65105eec3608d14f7324f37e005f714b70b899c08bab
3
+ size 7455061885
checkpoint-158000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b76b5fb716f18979eb2dbf80b017383e6a4bc77112e334800177a5ba0f3da1a
3
+ size 15024
checkpoint-158000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c0f0e75c7baefb2587805ec38529c726e1d888fffe4ec17fd7c90c444c1a4ae
3
+ size 15024
checkpoint-158000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a50574496c45c4721876b1982557d23ed5f83a8bfcc1250423e2cb041c168a1c
3
+ size 15024
checkpoint-158000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b16c4a49140afef2aff379e1a9121e657b33cb4e19163f666246c606862a0815
3
+ size 15024
checkpoint-158000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b01cbb678df92b7c5915930e224bb9d800cbeb11ea7d433741f5e97ae9d81a7
3
+ size 1064
checkpoint-158000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b03807553c71146e4452162f0f775ebaa46c6a1fd2c39bb57add55fb9915c45
3
+ size 24330100
checkpoint-158000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cc61ede1917ad0ae8f6fdbffae7132550860f2a9f141075be00fe887cd21966
3
+ size 6008
checkpoint-160000/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-160000/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-160000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ab83638062b54de219edd303795088a8647d0b6e08605fa2d511c04e6927fe5
3
+ size 4974360904
checkpoint-160000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b60752a01709e7c1650f15278fceba57257d30d717645c8fbf27803eff6ec56
3
+ size 3234250886
checkpoint-160000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-160000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:388d8dc0d18dfd24343a62bd46a6e285006c8b48250bcf74545fbae4b5d9c894
3
+ size 7455061885
checkpoint-160000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63640b9512a475c398bfea5f35bff6adfaac960416957fede9ae72f4883f465d
3
+ size 15024
checkpoint-160000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea59c6e349051dd34b1b8d00e4a18f7529ebdf68534de79785e1c936965e0be7
3
+ size 15024
checkpoint-160000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:528b080b6f9654fd80d7c098a2791d1984c72eebe0dd84cdbeb213d7b84a509e
3
+ size 15024
checkpoint-160000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8eef9bd685c4276cc8c5416266151e4ac771b1ec5b97a3adadb2d1be236b8a83
3
+ size 15024
checkpoint-160000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ac2b8c55d3fdf021e3e93dc337344834d29976ab064706df3d638988b57fc8a
3
+ size 1064
checkpoint-160000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:786dd39e2c447b0c81cb54ed80b039a18a42000e6faeddb4d2388f2e1e50e777
3
+ size 24639020
checkpoint-160000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cc61ede1917ad0ae8f6fdbffae7132550860f2a9f141075be00fe887cd21966
3
+ size 6008