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
a6b4217
·
verified ·
1 Parent(s): c345adf

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-188000/config.json +79 -0
  3. checkpoint-188000/generation_config.json +6 -0
  4. checkpoint-188000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-188000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-188000/model.safetensors.index.json +0 -0
  7. checkpoint-188000/optimizer.pt +3 -0
  8. checkpoint-188000/rng_state_0.pth +3 -0
  9. checkpoint-188000/rng_state_1.pth +3 -0
  10. checkpoint-188000/rng_state_2.pth +3 -0
  11. checkpoint-188000/rng_state_3.pth +3 -0
  12. checkpoint-188000/scheduler.pt +3 -0
  13. checkpoint-188000/trainer_state.json +3 -0
  14. checkpoint-188000/training_args.bin +3 -0
  15. checkpoint-190000/config.json +79 -0
  16. checkpoint-190000/generation_config.json +6 -0
  17. checkpoint-190000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-190000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-190000/model.safetensors.index.json +0 -0
  20. checkpoint-190000/optimizer.pt +3 -0
  21. checkpoint-190000/rng_state_0.pth +3 -0
  22. checkpoint-190000/rng_state_1.pth +3 -0
  23. checkpoint-190000/rng_state_2.pth +3 -0
  24. checkpoint-190000/rng_state_3.pth +3 -0
  25. checkpoint-190000/scheduler.pt +3 -0
  26. checkpoint-190000/trainer_state.json +3 -0
  27. checkpoint-190000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -82,3 +82,5 @@ 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
 
 
 
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
85
+ checkpoint-188000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
86
+ checkpoint-190000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-188000/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-188000/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-188000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3a5916c0e0008dddd894e92248b2eafa384d936a04dda59f893e358a68decfb
3
+ size 4974360904
checkpoint-188000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48d9e7ea89e00e52714aeaa09f4d407ba5c4bbf6a61de96fcf387a1478b6db9e
3
+ size 3234250886
checkpoint-188000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-188000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8fe337a35f7500b6a6e6748541d503efa3f0331b557cbdfc25390d7a753e96b
3
+ size 7455061885
checkpoint-188000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8c246e374dd84b3d34e4baa9189d0431852ef0ba857f85ba567d9f6606e1332
3
+ size 15024
checkpoint-188000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42678f0133f04dfb9b04978d2e3a46e7be930917d49323046e89c034277b424c
3
+ size 15024
checkpoint-188000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9865215dc8ed760e7524281512595e2826e0fe7ee89b1aece5c30eefc13d605f
3
+ size 15024
checkpoint-188000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f8c81df284fc296e6db4531af243286adb9cc9fe7dfec0079e9d0da74105601
3
+ size 15024
checkpoint-188000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b051ce66d2878716e65eee3b9533a63c03745206f415ce36f95ee8e07648ab8
3
+ size 1064
checkpoint-188000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5167f7bdd165977d4a4d30d67a433494e3b063f05035cc9aa7feb1026703c8a2
3
+ size 28962195
checkpoint-188000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f50737443324408b134388ddef0f6bf1281afaece998298b13a6a3af8e45d6e
3
+ size 6008
checkpoint-190000/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-190000/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-190000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28f9bac427caf81139ebd8610cd56f1e4b916f97068d175056e149095c24cff4
3
+ size 4974360904
checkpoint-190000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:993291b3c93066f07c2f046389e19cd75d49f40eb68c9c07a29767d578b46520
3
+ size 3234250886
checkpoint-190000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-190000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d79f35987f948ec5bb29264fc1c538239b4e1c1f25088135a65e1e93130ec89
3
+ size 7455061885
checkpoint-190000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a51de7c964102411887cf1c22c866474d67637a4ca5fb6bcfa359d6086a064f
3
+ size 15024
checkpoint-190000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c302546e994118f77c0289ca42fe1b176516388257027b0b9f6d1835a35f4c43
3
+ size 15024
checkpoint-190000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c31511c2a2b232e33df491799d1854debff078b0c5a177289172429fcc30d94e
3
+ size 15024
checkpoint-190000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:149cad62afc58f9a3754a72c67041d93b4ea88695e77190afa1801472c1e3ab0
3
+ size 15024
checkpoint-190000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4eb900306f8ea6f9b527d09b176c598275d9c2299f7a42e3e98f933ffd6c8a93
3
+ size 1064
checkpoint-190000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cd8411463284fca4c00a847771b76d8f50a158e2eb5991e4f276777777b7259
3
+ size 29270957
checkpoint-190000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f50737443324408b134388ddef0f6bf1281afaece998298b13a6a3af8e45d6e
3
+ size 6008