# RTM Latent Refinement (Fast Variant)

This model was trained with a reduced architecture (unknown parameters) compared to the original paper's baseline to speed up training.

- **Dataset**: CIFAR-10
- **W&B Logs**: Check the W&B project `rtm-latent-refinement` under the account `JerMa88`.
- **Source Code**: Included in the `code/` directory.