File size: 395 Bytes
15705a4 | 1 2 3 4 5 6 7 8 9 10 11 | # EMMeT
**E**mbedding **M**ixup for **Me**ta-**T**raining (EMMeT), a novel regularization technique that creates new episodic batches through an embedding shuffling and averaging regime.
Weights for the CONV4 and DINO-S models are provided for EMMeT evaluation.
For usage and instructions, please refer to the Github repository: https://github.com/Brikwerk/EMMeT
## Model Date
February 2023 |