{
  "accelerator_kwargs": {},
  "allow_tf32": true,
  "async_reward_computation": false,
  "exp_name": "ddpo",
  "log_with": "tensorboard",
  "logdir": "ll",
  "max_workers": 2,
  "mixed_precision": "fp16",
  "negative_prompts": "",
  "num_checkpoint_limit": 5,
  "num_epochs": 2,
  "per_prompt_stat_tracking": true,
  "per_prompt_stat_tracking_buffer_size": 4,
  "per_prompt_stat_tracking_min_count": 16,
  "project_kwargs": {
    "automatic_checkpoint_naming": true,
    "logging_dir": "./logs",
    "project_dir": "./save",
    "total_limit": 5
  },
  "resume_from": "",
  "run_name": "",
  "sample_batch_size": 1,
  "sample_eta": 1.0,
  "sample_guidance_scale": 5.0,
  "sample_num_batches_per_epoch": 2,
  "sample_num_steps": 5,
  "save_freq": 200,
  "seed": 0,
  "tracker_kwargs": {},
  "tracker_project_name": "stable_diffusion_training",
  "train_adam_beta1": 0.9,
  "train_adam_beta2": 0.999,
  "train_adam_epsilon": 1e-08,
  "train_adam_weight_decay": 0.0001,
  "train_adv_clip_max": 5,
  "train_batch_size": 1,
  "train_cfg": true,
  "train_clip_range": 0.0001,
  "train_gradient_accumulation_steps": 1,
  "train_learning_rate": 0.0003,
  "train_max_grad_norm": 1.0,
  "train_num_inner_epochs": 1,
  "train_timestep_fraction": 1.0,
  "train_use_8bit_adam": false
}