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
c345adf
·
verified ·
1 Parent(s): df9f629

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-178000/config.json +79 -0
  3. checkpoint-178000/generation_config.json +6 -0
  4. checkpoint-178000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-178000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-178000/model.safetensors.index.json +0 -0
  7. checkpoint-178000/optimizer.pt +3 -0
  8. checkpoint-178000/rng_state_0.pth +3 -0
  9. checkpoint-178000/rng_state_1.pth +3 -0
  10. checkpoint-178000/rng_state_2.pth +3 -0
  11. checkpoint-178000/rng_state_3.pth +3 -0
  12. checkpoint-178000/scheduler.pt +3 -0
  13. checkpoint-178000/trainer_state.json +3 -0
  14. checkpoint-178000/training_args.bin +3 -0
  15. checkpoint-180000/config.json +79 -0
  16. checkpoint-180000/generation_config.json +6 -0
  17. checkpoint-180000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-180000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-180000/model.safetensors.index.json +0 -0
  20. checkpoint-180000/optimizer.pt +3 -0
  21. checkpoint-180000/rng_state_0.pth +3 -0
  22. checkpoint-180000/rng_state_1.pth +3 -0
  23. checkpoint-180000/rng_state_2.pth +3 -0
  24. checkpoint-180000/rng_state_3.pth +3 -0
  25. checkpoint-180000/scheduler.pt +3 -0
  26. checkpoint-180000/trainer_state.json +3 -0
  27. checkpoint-180000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -80,3 +80,5 @@ 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
81
  checkpoint-168000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
82
  checkpoint-170000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
80
  checkpoint-160000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
81
  checkpoint-168000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
82
  checkpoint-170000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
83
+ checkpoint-178000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
84
+ checkpoint-180000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-178000/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-178000/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-178000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac41ed84c0eb8955d363b1496560d06734d3a60b0e416a337b59528848fa3ee0
3
+ size 4974360904
checkpoint-178000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:895765001e0e998230c15250c6814b6e300bfba2a1b10874e1da741e81e27039
3
+ size 3234250886
checkpoint-178000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-178000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ab4f3b045e1fa4ea3a46c125087eb01d81b2b4d8a65a4a810e2b6d3cc9057f3
3
+ size 7455061885
checkpoint-178000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50967afa4550278b9cf1f4b7bdba65b9c01d5bc1ce9e8cab7b15ae348a9c1e66
3
+ size 15024
checkpoint-178000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85baf3e179541cbf167279f2702af492f233c54ec8d775e89bf93fafff867a68
3
+ size 15024
checkpoint-178000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56169651496ae2475527183e0c17970538513de69da0fb60c01ccb9b869c1c02
3
+ size 15024
checkpoint-178000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6533d8807d22732ab2d335c833c314015f6f4cbf4b5fc2d7c105cd781d483951
3
+ size 15024
checkpoint-178000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:848fddaf212bbfa6108a1fd1e8015d769d8ebaf5f2ec413334ebe84e8584bb79
3
+ size 1064
checkpoint-178000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:593b565c42080067ab50e53b57bdabef0a737988102fd9402364040327c87b6b
3
+ size 27418231
checkpoint-178000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f46dc54151de8a5a5932c5109456be6be7aaf22580195da6c98f063137c1189b
3
+ size 6008
checkpoint-180000/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-180000/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-180000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5240cc090767988834eac4bb5e28637c55bd5a0c6fe0673c143204be74e2e2ad
3
+ size 4974360904
checkpoint-180000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eba3561b447e8dbd14356862a1e98854402bef476be783b19dc9c937b7eb50a1
3
+ size 3234250886
checkpoint-180000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-180000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86979e0254f541401e61d5e7a105dbe52dfbc8ab59fbd4f9d193fe880d0efa0c
3
+ size 7455061885
checkpoint-180000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:394d3d72c8942e167039ddb03ae6c119c45e176ebbc76e9c5cc24c881bd66f32
3
+ size 15024
checkpoint-180000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db97cbec8e71e6ccc4a678d1613b06f091109e02fe51bd311d3f70adfb02a9ff
3
+ size 15024
checkpoint-180000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4acca8ee63402f6eb992c1270d85bc3dcb5d9a6d12cae51b063346a770c75684
3
+ size 15024
checkpoint-180000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e93fef7d4e126230d2fa8ee9f96bba0eaeb38f21af0b24552b1568abf784b3c
3
+ size 15024
checkpoint-180000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7da6a4d349d318b433760ce537d4070978793414b6754bd1cf74b623a785c9b7
3
+ size 1064
checkpoint-180000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ce1e286d30ab883a09c58c8daf2dcaa486827cddbb267040531f5967c47e268
3
+ size 27726975
checkpoint-180000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f46dc54151de8a5a5932c5109456be6be7aaf22580195da6c98f063137c1189b
3
+ size 6008