kyujinpy commited on
Commit
bde9e93
β€’
1 Parent(s): 4101ecc

Upload 4 files

Browse files
config.yaml ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ pretrained_model_path: ./checkpoints/Bingsu/my-korean-stable-diffusion-v1-5
2
+ output_dir: outputs/bear_guitar
3
+ train_data:
4
+ video_path: data/man-surfing.mp4
5
+ prompt: A man is surfing
6
+ n_sample_frames: 24
7
+ width: 512
8
+ height: 512
9
+ sample_start_idx: 0
10
+ sample_frame_rate: 2
11
+ validation_data:
12
+ prompts:
13
+ - λ―Έν‚€λ§ˆμš°μŠ€κ°€ μ„œν•‘μ„ 타고 μžˆμŠ΅λ‹ˆλ‹€
14
+ - ν•œ μ—¬μžκ°€ μ„œν•‘μ„ 타고 μžˆμŠ΅λ‹ˆλ‹€
15
+ - 흰색 μ˜·μ„ μž…μ€ λ‚¨μžκ°€ λ°”λ‹€λ₯Ό κ±·κ³  μžˆμŠ΅λ‹ˆλ‹€
16
+ video_length: 24
17
+ width: 512
18
+ height: 512
19
+ num_inference_steps: 20
20
+ guidance_scale: 12.5
21
+ use_inv_latent: true
22
+ num_inv_steps: 50
23
+ validation_steps: 100
24
+ trainable_modules:
25
+ - attn1.to_q
26
+ - attn2.to_q
27
+ - attn_temp
28
+ train_batch_size: 1
29
+ max_train_steps: 400
30
+ learning_rate: 3.0e-05
31
+ scale_lr: false
32
+ lr_scheduler: constant
33
+ lr_warmup_steps: 0
34
+ adam_beta1: 0.9
35
+ adam_beta2: 0.999
36
+ adam_weight_decay: 0.01
37
+ adam_epsilon: 1.0e-08
38
+ max_grad_norm: 1.0
39
+ gradient_accumulation_steps: 1
40
+ gradient_checkpointing: true
41
+ checkpointing_steps: 1000
42
+ resume_from_checkpoint: null
43
+ mixed_precision: fp16
44
+ use_8bit_adam: false
45
+ enable_xformers_memory_efficient_attention: true
46
+ seed: 33
model_index.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "TuneAVideoPipeline",
3
+ "_diffusers_version": "0.11.1",
4
+ "scheduler": [
5
+ "diffusers",
6
+ "PNDMScheduler"
7
+ ],
8
+ "text_encoder": [
9
+ "transformers",
10
+ "CLIPTextModel"
11
+ ],
12
+ "tokenizer": [
13
+ "transformers",
14
+ "CLIPTokenizer"
15
+ ],
16
+ "unet": [
17
+ "models",
18
+ "UNet3DConditionModel"
19
+ ],
20
+ "vae": [
21
+ "diffusers",
22
+ "AutoencoderKL"
23
+ ]
24
+ }
vae/config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "AutoencoderKL",
3
+ "_diffusers_version": "0.11.1",
4
+ "_name_or_path": "./checkpoints/Bingsu/my-korean-stable-diffusion-v1-5",
5
+ "act_fn": "silu",
6
+ "block_out_channels": [
7
+ 128,
8
+ 256,
9
+ 512,
10
+ 512
11
+ ],
12
+ "down_block_types": [
13
+ "DownEncoderBlock2D",
14
+ "DownEncoderBlock2D",
15
+ "DownEncoderBlock2D",
16
+ "DownEncoderBlock2D"
17
+ ],
18
+ "in_channels": 3,
19
+ "latent_channels": 4,
20
+ "layers_per_block": 2,
21
+ "norm_num_groups": 32,
22
+ "out_channels": 3,
23
+ "sample_size": 512,
24
+ "scaling_factor": 0.18215,
25
+ "up_block_types": [
26
+ "UpDecoderBlock2D",
27
+ "UpDecoderBlock2D",
28
+ "UpDecoderBlock2D",
29
+ "UpDecoderBlock2D"
30
+ ]
31
+ }
vae/diffusion_pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:081334a8acb2a2eecbee377823b2819669d45f56c593159a20d13754c96571fa
3
+ size 167407601