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
973e899
·
verified ·
1 Parent(s): 29466c8

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-278000/config.json +79 -0
  3. checkpoint-278000/generation_config.json +6 -0
  4. checkpoint-278000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-278000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-278000/model.safetensors.index.json +0 -0
  7. checkpoint-278000/optimizer.pt +3 -0
  8. checkpoint-278000/rng_state_0.pth +3 -0
  9. checkpoint-278000/rng_state_1.pth +3 -0
  10. checkpoint-278000/rng_state_2.pth +3 -0
  11. checkpoint-278000/rng_state_3.pth +3 -0
  12. checkpoint-278000/scheduler.pt +3 -0
  13. checkpoint-278000/trainer_state.json +3 -0
  14. checkpoint-278000/training_args.bin +3 -0
  15. checkpoint-280000/config.json +79 -0
  16. checkpoint-280000/generation_config.json +6 -0
  17. checkpoint-280000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-280000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-280000/model.safetensors.index.json +0 -0
  20. checkpoint-280000/optimizer.pt +3 -0
  21. checkpoint-280000/rng_state_0.pth +3 -0
  22. checkpoint-280000/rng_state_1.pth +3 -0
  23. checkpoint-280000/rng_state_2.pth +3 -0
  24. checkpoint-280000/rng_state_3.pth +3 -0
  25. checkpoint-280000/scheduler.pt +3 -0
  26. checkpoint-280000/trainer_state.json +3 -0
  27. checkpoint-280000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -115,3 +115,5 @@ checkpoint-258000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
115
  checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
116
  checkpoint-268000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
117
  checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
115
  checkpoint-260000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
116
  checkpoint-268000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
117
  checkpoint-270000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
118
+ checkpoint-278000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
119
+ checkpoint-280000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-278000/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-278000/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-278000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1bfce2bed62294e0c037f907924eff69b593c111693f50bf300fb8dd9ae821d6
3
+ size 4974360904
checkpoint-278000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1d893c8aacd2a529922fa92bfa96b8c80ed90d6fe73fd06df3ffe7ef0b15991
3
+ size 3234250886
checkpoint-278000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-278000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:028af008f74fb38be4b7597d5165fc955e2fa1d0fc08ba10b88d410c19f56583
3
+ size 7455061885
checkpoint-278000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de0d074b50d11423c08d1a8141f89878b35fe87e77961d27b24b89a1ac63da21
3
+ size 15024
checkpoint-278000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7789540e42f333887e09f76640252e98e3b6250f39519a1f1b3a5767d39e6796
3
+ size 15024
checkpoint-278000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f00039671fea2bee667a227dbcdb580f793d85b1a0e1886a9595084998c5aed
3
+ size 15024
checkpoint-278000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e05dcd4a80b2379f498598122d84738b243999658e162fe465a6f90e4c895e5
3
+ size 15024
checkpoint-278000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5354b7730049075219c88230e505c69768ad700259243d2409644df4c958a983
3
+ size 1064
checkpoint-278000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93dd4d321c1b646318ee3f63b92e7d290aaa3315334f8268a12082511a92a612
3
+ size 42845011
checkpoint-278000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98043351e3a3298e22fcf5cc6297f24975d90e5761fb83c299fde67035ff9689
3
+ size 6072
checkpoint-280000/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-280000/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-280000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9959e63f6c7d0c1b31837565d0fd0cef2dde70950df057a4e9c1da58482f8ca
3
+ size 4974360904
checkpoint-280000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3217a5c7ce9e05e47f31336cfd553df510fd7e944e2f63ead8d4a404667ca18
3
+ size 3234250886
checkpoint-280000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-280000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f1716333de0dd3a3ca012bfcdde8729485f7d15ce1e46d1d1c93989635b389d
3
+ size 7455061885
checkpoint-280000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27c52d8ebdd680f730a18d8ad5579450661370e34cad1199b81081d700b94cfb
3
+ size 15024
checkpoint-280000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f46535b489b1ce8330a0108fd2323a49a026f7d70bd1ef9f7876501ce87255ff
3
+ size 15024
checkpoint-280000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d617144125e02820e36ac955eacb8a10305f26e5a5093627347c9a277eec3d90
3
+ size 15024
checkpoint-280000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c251d1095d5d9dd448d97f3230eff994241e971ef1c5d5b2900f952b95a5863
3
+ size 15024
checkpoint-280000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd6eb430790fcda079f12a19f414bf8ecb9ffe09b988dbc1fe0b2a4bab209523
3
+ size 1064
checkpoint-280000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ca423c710c9f8d3423751c7524c082fea1bc855bf7e41cdc47660cbdb3f88e2
3
+ size 43153350
checkpoint-280000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98043351e3a3298e22fcf5cc6297f24975d90e5761fb83c299fde67035ff9689
3
+ size 6072