# Wan14B DMD: freeze the distilled Full Generator and train Predictor-v4.
trainer: predictor_v4_dmd
distribution_loss: dmd
mixed_precision: true
seed: 0

generator_fsdp_wrap_strategy: size
real_score_fsdp_wrap_strategy: size
fake_score_fsdp_wrap_strategy: size
text_encoder_fsdp_wrap_strategy: size
text_encoder_cpu_offload: false
real_name: Wan2.1-T2V-14B
fake_name: Wan2.1-T2V-1.3B
sharding_strategy: hybrid_full

denoising_step_list: [1000, 750, 500, 250]
warp_denoising_step: true
num_train_timestep: 1000
timestep_shift: 5.0
guidance_scale: 3.0
denoising_loss_type: flow
gradient_checkpointing: true
num_frame_per_block: 3
num_training_frames: 21
same_step_across_blocks: true
last_step_only: false
independent_first_frame: false
context_noise: 0
ts_schedule: false
min_score_timestep: 0
i2v: false
load_raw_video: false
batch_size: 1
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'
image_or_video_shape: [1, 21, 16, 60, 104]
model_kwargs:
  timestep_shift: 5.0

predictor_v4_dmd:
  training_mode: predictor_only
  expected_world_size: 8
  data_path: prompts/vidprom_filtered_extended.txt
  output_dir: /mnt/local_nvme/zoubin/cz/self_forcing_predictor_v4_dmd_predictor_only

  full_checkpoint: checkpoints/self_forcing_dmd.pt
  full_checkpoint_key: generator_ema
  strict_full_load: true
  stage1_training_state: /mnt/local_nvme/zoubin/cz/self_forcing_predictor_v4_training_1000_seed0/training_latest.pt
  stage1_expected_step: 2000
  source_block_ids: [1, 28]

  max_student_steps: 2000
  target_predictor_updates: 2000
  critic_updates_per_student: 5
  predictor_fusion_lr: 1.0e-5
  predictor_blocks_lr: 1.0e-6
  full_lr: 0.0
  fake_score_lr: 4.0e-7
  student_beta1: 0.0
  student_beta2: 0.999
  critic_beta1: 0.0
  critic_beta2: 0.999
  weight_decay: 0.01
  warmup_steps: 100
  predictor_grad_clip: 1.0
  full_grad_clip: 1.0
  critic_grad_clip: 10.0

  ema_decay: 0.99
  ema_start_step: 200
  forced_exit_step: null
  num_workers: 0
  pin_memory: false
  log_every: 1
  save_every: 200
  keep_checkpoints: 2

  use_swanlab: true
  swanlab_project: Self-Forcing-Predictor-v4
  swanlab_workspace: null
  swanlab_experiment: predictor-v4-wan14b-dmd-predictor-only-p1p2p3
  swanlab_description: Wan14B DMD with frozen Full Generator and random P1/P2/P3 Predictor exits
  swanlab_tags:
    - predictor-v4
    - wan14b-dmd
    - predictor-only
    - random-exit
    - p1-p2-p3
    - 8gpu
  swanlab_mode: cloud