Text-to-Video
English
File size: 1,012 Bytes
310cdb3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
# dynamic sample + change rope
dynamic_sample_frames: True
change_rope: True
max_context_frames: 6 
multi_caption: True

# model
model_name: Wan2.1-T2V-1.3B
resume_ckpt: None

denoising_step_list:
- 1000
- 740
- 500
- 260
warp_denoising_step: True # need to remove - 0 in denoising_step_list if warp_denoising_step is true
num_frame_per_block: 3

seed: 0
ts_schedule: false
num_train_timestep: 1000
timestep_shift: 8.0
guidance_scale: 3.0
mixed_precision: true
sharding_strategy: hybrid_full

# Data
data_path: None
batch_size: 1

# log 
log_iters: 50
max_iters: 30000
negative_prompt: '色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走'

output_folder: None