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
6470fc2
·
verified ·
1 Parent(s): 026e4b9

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-88000/config.json +79 -0
  3. checkpoint-88000/generation_config.json +6 -0
  4. checkpoint-88000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-88000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-88000/model.safetensors.index.json +0 -0
  7. checkpoint-88000/optimizer.pt +3 -0
  8. checkpoint-88000/rng_state_0.pth +3 -0
  9. checkpoint-88000/rng_state_1.pth +3 -0
  10. checkpoint-88000/rng_state_2.pth +3 -0
  11. checkpoint-88000/rng_state_3.pth +3 -0
  12. checkpoint-88000/scheduler.pt +3 -0
  13. checkpoint-88000/trainer_state.json +3 -0
  14. checkpoint-88000/training_args.bin +3 -0
  15. checkpoint-90000/config.json +79 -0
  16. checkpoint-90000/generation_config.json +6 -0
  17. checkpoint-90000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-90000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-90000/model.safetensors.index.json +0 -0
  20. checkpoint-90000/optimizer.pt +3 -0
  21. checkpoint-90000/rng_state_0.pth +3 -0
  22. checkpoint-90000/rng_state_1.pth +3 -0
  23. checkpoint-90000/rng_state_2.pth +3 -0
  24. checkpoint-90000/rng_state_3.pth +3 -0
  25. checkpoint-90000/scheduler.pt +3 -0
  26. checkpoint-90000/trainer_state.json +3 -0
  27. checkpoint-90000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -36,3 +36,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
36
  checkpoint-70000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
37
  checkpoint-78000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
38
  checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
36
  checkpoint-70000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
37
  checkpoint-78000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
38
  checkpoint-80000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
39
+ checkpoint-88000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
40
+ checkpoint-90000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-88000/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-88000/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-88000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1425360a5f5ab373b6d92cfd535e44c021ef0df5c9d4c47b92b56cb5d5150d0f
3
+ size 4974360904
checkpoint-88000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40d48d32bb97a206bd1ef200ea5dba027c310e497e308f5577adb0325384ee65
3
+ size 3234250886
checkpoint-88000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-88000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6966773d38fbd7fcbf3eeaf293790bd6d5259c6fc67e64621a1128cf83d1ea97
3
+ size 7455061885
checkpoint-88000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35bf0328951b6b75549f97e3afa97e4eb3d85ab75c6d621a52ebce4b4a7a1eba
3
+ size 15024
checkpoint-88000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c5c0d8d0568f14f066c13227108895c9bd9dc4f1c45a165010a4c2edd768736
3
+ size 15024
checkpoint-88000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e872c1ebf0b5c37c8ae6cf20384b1a5e942df9168fd77e2b9ad9326ffe392e0
3
+ size 15024
checkpoint-88000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bfbe0a6289b9deb554fed3ef0910b03e4ee77bd9ce5a12fffdfb4878816aa0af
3
+ size 15024
checkpoint-88000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c45458450455ee9cd92f20a9e24aa01d17791de2bc975e46aba6d63edea82398
3
+ size 1064
checkpoint-88000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c49218932f38cdc9795a68c0f35808688acf3f25617e6d5802a1a4acb979ad20
3
+ size 13533176
checkpoint-88000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:becdce52c48ee8757b4b723d61dba9b0a47d76a2de45d9f0f267ac150b64c968
3
+ size 6008
checkpoint-90000/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-90000/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-90000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c388e1de3dc07cc5201549c2075243415be8c1ffa58098185b93d3b42674ab02
3
+ size 4974360904
checkpoint-90000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa401930ff38f3fc811485d4138b59b1e5942b1208c6f7e5c207838aca394bf8
3
+ size 3234250886
checkpoint-90000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-90000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80abc21af0abd741c0ecd1146dbb09873649e113ae48a8931d426c53de7eb1cc
3
+ size 7455061885
checkpoint-90000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:457d17594cd9581d36f82acc6f71d91aaa0b9835fd6301610b62968fa3972fb3
3
+ size 15024
checkpoint-90000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a2a0e1a2503d633b4b7e1a45fd7446920e4fb2841174e9c866446c4c583b566
3
+ size 15024
checkpoint-90000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83c2347193350d389d0a661600a67e7fef0a30a618dd44ae906e58b13b3b6ea3
3
+ size 15024
checkpoint-90000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6b4429d283ce62ea86c5fec328cb5eb6d6d9d0a4815ffd254ff06bc8e351360
3
+ size 15024
checkpoint-90000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f17fac99609a9eef75b7ab41b9b39e47bf03a67a9c70b8a6c43cc3eb54f185ac
3
+ size 1064
checkpoint-90000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e1b6ad7a864029e0e2477560828a582f10b575ff442466d554ee3ee434399de
3
+ size 13839937
checkpoint-90000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:becdce52c48ee8757b4b723d61dba9b0a47d76a2de45d9f0f267ac150b64c968
3
+ size 6008