| - run_id=bs128_ld512_ts200_big_model | |
| - trainer.strategy=auto | |
| - devices=1 | |
| - machine.num_workers=8 | |
| - experiment=final-model | |
| - machine.batch_size=128 | |
| - model.diff_params.num_train_timesteps=200 | |
| - model.latent_dim=512 | |
| - model.ff_size=2048 | |
| - model.num_layers=8 | |
| - model.num_head=8 | |