self-forcing / configs /self_forcing_dmd.yaml
multimodalart's picture
Upload 80 files
0fd2f06 verified
generator_ckpt: checkpoints/ode_init.pt
generator_fsdp_wrap_strategy: size
real_score_fsdp_wrap_strategy: size
fake_score_fsdp_wrap_strategy: size
real_name: Wan2.1-T2V-14B
text_encoder_fsdp_wrap_strategy: size
denoising_step_list:
- 1000
- 750
- 500
- 250
warp_denoising_step: true # need to remove - 0 in denoising_step_list if warp_denoising_step is true
ts_schedule: false
num_train_timestep: 1000
timestep_shift: 5.0
guidance_scale: 3.0
denoising_loss_type: flow
mixed_precision: true
seed: 0
wandb_host: WANDB_HOST
wandb_key: WANDB_KEY
wandb_entity: WANDB_ENTITY
wandb_project: WANDB_PROJECT
sharding_strategy: hybrid_full
lr: 2.0e-06
lr_critic: 4.0e-07
beta1: 0.0
beta2: 0.999
beta1_critic: 0.0
beta2_critic: 0.999
data_path: prompts/vidprom_filtered_extended.txt
batch_size: 1
ema_weight: 0.99
ema_start_step: 200
total_batch_size: 64
log_iters: 50
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'
dfake_gen_update_ratio: 5
image_or_video_shape:
- 1
- 21
- 16
- 60
- 104
distribution_loss: dmd
trainer: score_distillation
gradient_checkpointing: true
num_frame_per_block: 3
load_raw_video: false
model_kwargs:
timestep_shift: 5.0