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
8926b29
·
verified ·
1 Parent(s): 2ad6701

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-258000/config.json +79 -0
  3. checkpoint-258000/generation_config.json +6 -0
  4. checkpoint-258000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-258000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-258000/model.safetensors.index.json +0 -0
  7. checkpoint-258000/optimizer.pt +3 -0
  8. checkpoint-258000/rng_state_0.pth +3 -0
  9. checkpoint-258000/rng_state_1.pth +3 -0
  10. checkpoint-258000/rng_state_2.pth +3 -0
  11. checkpoint-258000/rng_state_3.pth +3 -0
  12. checkpoint-258000/scheduler.pt +3 -0
  13. checkpoint-258000/trainer_state.json +3 -0
  14. checkpoint-258000/training_args.bin +3 -0
  15. checkpoint-260000/config.json +79 -0
  16. checkpoint-260000/generation_config.json +6 -0
  17. checkpoint-260000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-260000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-260000/model.safetensors.index.json +0 -0
  20. checkpoint-260000/optimizer.pt +3 -0
  21. checkpoint-260000/rng_state_0.pth +3 -0
  22. checkpoint-260000/rng_state_1.pth +3 -0
  23. checkpoint-260000/rng_state_2.pth +3 -0
  24. checkpoint-260000/rng_state_3.pth +3 -0
  25. checkpoint-260000/scheduler.pt +3 -0
  26. checkpoint-260000/trainer_state.json +3 -0
  27. checkpoint-260000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -111,3 +111,5 @@ checkpoint-238000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
111
  checkpoint-240000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
112
  checkpoint-248000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
113
  checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
111
  checkpoint-240000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
112
  checkpoint-248000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
113
  checkpoint-250000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
114
+ 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
checkpoint-258000/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-258000/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-258000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2a5768a22e39da0f9f91c93d2caadabdd9606559d598d7dbf7b014fb11dbd0c
3
+ size 4974360904
checkpoint-258000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:070b0f3f10353b89c9dae5a2dbce7a80d93521aa3bcd6fbc90de2752221a9195
3
+ size 3234250886
checkpoint-258000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-258000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c07f0a8be6fd6a29aa5890a531b321d1c2f1fe4719a62a78f5c9a8fc893bcaa3
3
+ size 7455061885
checkpoint-258000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f816c6a53eb3391fe9c008769ae4dfd83ea03677530370e397facb7b044c38ab
3
+ size 15024
checkpoint-258000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f91b19e0e528a2fe503a0f6a532427927aee5d0e64f9949882c9fa6143f7498
3
+ size 15024
checkpoint-258000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45a2d1309eeab15c79af2fcfbb4b157ce8b961fc34e7c69c73fc82b3223c854a
3
+ size 15024
checkpoint-258000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b1a60c47173e67053733d5e63f8da1d2f1c5d0d73fdf12bf78d1c9f8d260f2e
3
+ size 15024
checkpoint-258000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55980c72aa8e8f23a49147634f3062c619486c1d92eeb05f167c87f14e69e9af
3
+ size 1064
checkpoint-258000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3e6450f7e8f35e274ebd9fe979e53f4092dd9d6f52b4438f9d775a1fa8c66cc
3
+ size 39762135
checkpoint-258000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4e09d9b51390c948bb26e3db38a4574be451dd5d351c510d5e402f4aad439ff
3
+ size 6072
checkpoint-260000/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-260000/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-260000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6c79aa21699bd672cd2bac4e45f7fca138c3f2034fedecd576d5072a9e00ef5
3
+ size 4974360904
checkpoint-260000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdccbb54034d03b289c4e4496961cb00fe50145badde40d3dd1bf851f873956f
3
+ size 3234250886
checkpoint-260000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-260000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a3ead0935e522fc27abe06d5af49629f3483604217ba1454fc867f683ee1e19
3
+ size 7455061885
checkpoint-260000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e62cf6c62e936db256d37ae7984490b0c1069029716dfa75d95cd69df8eabe02
3
+ size 15024
checkpoint-260000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7423f5849cdb6eee21997b1695b390393eb53eb92cee5a715aab53ba8eed45c
3
+ size 15024
checkpoint-260000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6bcdaa082630eff485fdf46d1f197933482b9a9294e54eeeaa613a58b47d4d8d
3
+ size 15024
checkpoint-260000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cabe61a7e45e174b32713b1208d310c69dba3cbb341ce744cd9d7e1d76664b23
3
+ size 15024
checkpoint-260000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2056bcd62c0c4c840a09c0e3f022709245d8b64def7eaad36decf9f8b6fb8427
3
+ size 1064
checkpoint-260000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d27ba0469300a395d98b31f9df2c0ddb6f61dd1e4e7fe58bf50a71dd04a0403
3
+ size 40070369
checkpoint-260000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4e09d9b51390c948bb26e3db38a4574be451dd5d351c510d5e402f4aad439ff
3
+ size 6072