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
df9f629
·
verified ·
1 Parent(s): 6aea61a

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-168000/config.json +79 -0
  3. checkpoint-168000/generation_config.json +6 -0
  4. checkpoint-168000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-168000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-168000/model.safetensors.index.json +0 -0
  7. checkpoint-168000/optimizer.pt +3 -0
  8. checkpoint-168000/rng_state_0.pth +3 -0
  9. checkpoint-168000/rng_state_1.pth +3 -0
  10. checkpoint-168000/rng_state_2.pth +3 -0
  11. checkpoint-168000/rng_state_3.pth +3 -0
  12. checkpoint-168000/scheduler.pt +3 -0
  13. checkpoint-168000/trainer_state.json +3 -0
  14. checkpoint-168000/training_args.bin +3 -0
  15. checkpoint-170000/config.json +79 -0
  16. checkpoint-170000/generation_config.json +6 -0
  17. checkpoint-170000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-170000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-170000/model.safetensors.index.json +0 -0
  20. checkpoint-170000/optimizer.pt +3 -0
  21. checkpoint-170000/rng_state_0.pth +3 -0
  22. checkpoint-170000/rng_state_1.pth +3 -0
  23. checkpoint-170000/rng_state_2.pth +3 -0
  24. checkpoint-170000/rng_state_3.pth +3 -0
  25. checkpoint-170000/scheduler.pt +3 -0
  26. checkpoint-170000/trainer_state.json +3 -0
  27. checkpoint-170000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -78,3 +78,5 @@ 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
 
 
 
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
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
checkpoint-168000/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-168000/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-168000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d19c934d2ced00b78b9841cbd581dfaa855a65822f5588647761842fbc045d2
3
+ size 4974360904
checkpoint-168000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e556e22532375109b5f4cf9d5b2082e6d0962c7d49220d52291a450b46d0496
3
+ size 3234250886
checkpoint-168000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-168000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c08b3b9242b9b9ba9e2b2b14b040d3ab426628dbd857879c34f86da99c6a0630
3
+ size 7455061885
checkpoint-168000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:463d89f5072f0752d289d471959f9cc8acfa812a70f82903026bdbd3a00e9505
3
+ size 15024
checkpoint-168000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18d86e95527baeb884c95d167e06c2e39c61a71a0255425afaad8529e753db20
3
+ size 15024
checkpoint-168000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab19eec0e2f98582f62e0b1aecb5bc61810e0dc47f8b0d69399f449ed2b525cc
3
+ size 15024
checkpoint-168000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d0a76de671dd87295b1b85a5850afc82cbeecea81b6f979708bbe07fb17da60
3
+ size 15024
checkpoint-168000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f27035c7b2b1e1dbc4c9dafc26d00214857b52178d2506157f12f4c410a289e1
3
+ size 1064
checkpoint-168000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27938fcac24c8b4aef48d09e02ac3f2019d5b14fa14dd206acad4b16d88dd079
3
+ size 25874196
checkpoint-168000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e49e7611c17101fe11604c213048e91e00dbc960bc20f2e2ea68d77fc70fc37
3
+ size 6008
checkpoint-170000/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-170000/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-170000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f62143cd882f635d1f75be302259f460fc06d85f24407facfcf081d318cd512
3
+ size 4974360904
checkpoint-170000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d2de0e116e5482ce2e29598fe416377fbb4757403e46ad4f1c7e8dca20dae15
3
+ size 3234250886
checkpoint-170000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-170000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a8d00516325458f0471e4d3af34ef24890899692e027c256cc0ebfc9ed17e4a
3
+ size 7455061885
checkpoint-170000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2e72ba1264a04ec41814fdc13d093ac3e4ea1a3b92c9d66a5b3d293cf21646b
3
+ size 15024
checkpoint-170000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f35f26edc5aa70a2bf13f7e6170d84e047da6838849576dbbc7060aba10986b
3
+ size 15024
checkpoint-170000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6290343ad2f7e4a1c17c369af101b05833fa8cfca21d45ae9ff145ce80cc8f0
3
+ size 15024
checkpoint-170000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b615b4d8bb63c85e9a65428980a6acc90ac92cb3d6e7304de75f57019d95f50
3
+ size 15024
checkpoint-170000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:230cfffa87b85649ed86bd4b446d61bf13735f62ba43398a798cf217a49f21e8
3
+ size 1064
checkpoint-170000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15003e373deea6b1a03972b9b2a6cefc1af212bd0221209d233b082fdef7f5a9
3
+ size 26183146
checkpoint-170000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e49e7611c17101fe11604c213048e91e00dbc960bc20f2e2ea68d77fc70fc37
3
+ size 6008