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
9fe1a97
·
verified ·
1 Parent(s): 1f298b5

Upload folder using huggingface_hub

Browse files
Files changed (27) hide show
  1. .gitattributes +2 -0
  2. checkpoint-118000/config.json +79 -0
  3. checkpoint-118000/generation_config.json +6 -0
  4. checkpoint-118000/model-00001-of-00002.safetensors +3 -0
  5. checkpoint-118000/model-00002-of-00002.safetensors +3 -0
  6. checkpoint-118000/model.safetensors.index.json +0 -0
  7. checkpoint-118000/optimizer.pt +3 -0
  8. checkpoint-118000/rng_state_0.pth +3 -0
  9. checkpoint-118000/rng_state_1.pth +3 -0
  10. checkpoint-118000/rng_state_2.pth +3 -0
  11. checkpoint-118000/rng_state_3.pth +3 -0
  12. checkpoint-118000/scheduler.pt +3 -0
  13. checkpoint-118000/trainer_state.json +3 -0
  14. checkpoint-118000/training_args.bin +3 -0
  15. checkpoint-120000/config.json +79 -0
  16. checkpoint-120000/generation_config.json +6 -0
  17. checkpoint-120000/model-00001-of-00002.safetensors +3 -0
  18. checkpoint-120000/model-00002-of-00002.safetensors +3 -0
  19. checkpoint-120000/model.safetensors.index.json +0 -0
  20. checkpoint-120000/optimizer.pt +3 -0
  21. checkpoint-120000/rng_state_0.pth +3 -0
  22. checkpoint-120000/rng_state_1.pth +3 -0
  23. checkpoint-120000/rng_state_2.pth +3 -0
  24. checkpoint-120000/rng_state_3.pth +3 -0
  25. checkpoint-120000/scheduler.pt +3 -0
  26. checkpoint-120000/trainer_state.json +3 -0
  27. checkpoint-120000/training_args.bin +3 -0
.gitattributes CHANGED
@@ -42,3 +42,5 @@ checkpoint-100000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
42
  checkpoint-98000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
43
  checkpoint-108000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
44
  checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
 
 
 
42
  checkpoint-98000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
43
  checkpoint-108000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
44
  checkpoint-110000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
45
+ checkpoint-118000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
46
+ checkpoint-120000/trainer_state.json filter=lfs diff=lfs merge=lfs -text
checkpoint-118000/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-118000/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-118000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db7875fb34a5197f405830002c7bdf2ce6b2bbfc0357520bd876bd64087b49ca
3
+ size 4974360904
checkpoint-118000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad80013b9c96425728eca52d4b4d04991804874cdef7f46ae7eac303077ff7c3
3
+ size 3234250886
checkpoint-118000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-118000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be33114f9e7e6032ba55302de3abdd2ed812ba2fc37c5caefb9f499fbd3709d4
3
+ size 7455061885
checkpoint-118000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a854ae87bd993fc436cb189051a50278e21955e7cbd034d921cc08c7eb2b3df2
3
+ size 15024
checkpoint-118000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e43eb92002d63012ce8e9392c4aa1941a77c23e51c14e7c7d494316645f60878
3
+ size 15024
checkpoint-118000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a143deacc5dba1bdecda0cec563dc02c1301e33917e8705b383b7748d09b108
3
+ size 15024
checkpoint-118000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed6d0eac4f3df463fec3f96c8b5b7552fd9ecbe7bc478d016a3648e1dabc5ce6
3
+ size 15024
checkpoint-118000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9877127b547b3526cbc0b0d6c26d19f1d21967a863081efad927166983806b1
3
+ size 1064
checkpoint-118000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74f00e1e6d582b4f346a45d1757f1533b4cc4e1d7de1cc7abc0c2ce8a5661fbd
3
+ size 18155474
checkpoint-118000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9ecef7ebcb8f8fa46e23d105f507a3ae35db2f8e725c815543adc3702c489ab
3
+ size 6008
checkpoint-120000/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-120000/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-120000/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff96827d448f1ae728e3f39e7896ee8e7065303ee8b275e5ace0c8cd9afc63b5
3
+ size 4974360904
checkpoint-120000/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12abf74c7803408891707fd9d4af0793bea0b3ccdd7e1dbae1dd4b4364361c90
3
+ size 3234250886
checkpoint-120000/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-120000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3cbf366e48d11c08a02b5b53ec3493ab076f1b81befdb1738d902b94a0dd245
3
+ size 7455061885
checkpoint-120000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6fe504bcbaab673d50eb0e43b13e2740dee97ff599fa87482b338820fc72683
3
+ size 15024
checkpoint-120000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca321b3335a819a9e456167a425f1310120fa18b8e7b21e8241b1004f43fb499
3
+ size 15024
checkpoint-120000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b49384140b8b805059fba18d1cc796b3b75ee6cece89c770ec950d38956adb77
3
+ size 15024
checkpoint-120000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b09c48afe400b10e160080fabf4728f3d43b39a3d8feb943777851f629ade1a7
3
+ size 15024
checkpoint-120000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:336307835178d65b3db354b19d0080ec0ecb705b1ebcd556dd0d62657bcd5f75
3
+ size 1064
checkpoint-120000/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7218827f37663c37b19292c8456b6dbefbb19e50d7a5d1824abe108cc0722a09
3
+ size 18464281
checkpoint-120000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9ecef7ebcb8f8fa46e23d105f507a3ae35db2f8e725c815543adc3702c489ab
3
+ size 6008