| adam_beta1: 0.95 | |
| adam_beta2: 0.999 | |
| adam_epsilon: 1.0e-08 | |
| adam_weight_decay: 1.0e-06 | |
| cache_dir: null | |
| center_crop: true | |
| checkpointing_steps: 500 | |
| checkpoints_total_limit: null | |
| dataloader_num_workers: 0 | |
| dataset_config_name: null | |
| dataset_name: cifar10 | |
| ema_inv_gamma: 1.0 | |
| ema_max_decay: 0.9999 | |
| ema_power: 0.75 | |
| eval_batch_size: 16 | |
| gradient_accumulation_steps: 1 | |
| hub_model_id: null | |
| hub_private_repo: false | |
| hub_token: null | |
| learning_rate: 0.0001 | |
| local_rank: -1 | |
| logger: tensorboard | |
| logging_dir: logs | |
| lr_scheduler: cosine | |
| lr_warmup_steps: 500 | |
| mixed_precision: 'no' | |
| model_config_name_or_path: null | |
| num_epochs: 100 | |
| output_dir: cifar10-32 | |
| overwrite_output_dir: false | |
| pretrained_teacher_model_name_or_path: google/ddpm-cifar10-32 | |
| push_to_hub: true | |
| random_flip: true | |
| resolution: 32 | |
| resume_from_checkpoint: null | |
| save_images_epochs: 10 | |
| save_model_epochs: 10 | |
| train_batch_size: 16 | |
| train_data_dir: null | |