metadata
license: cc-by-4.0
tags:
- co-attention
- deit
- polynomial
- retrieval
- rmsnorm
- rmsprop
- standard
- swish
- tiny
- xavier
model_562153981_deit_tiny.py
Model Overview
A tiny-scale implementation of the deit architecture, built for retrieval tasks.
Architecture
- Architecture: deit
- Scale: tiny
- Attention: standard
- Fusion strategy: co attention
- Task head: retrieval
- Activation: swish
- Normalization: rmsnorm
- Initialization: xavier
Training
- Optimizer: rmsprop
- LR scheduler: polynomial
Files
model_562153981_deit_tiny.py— main artifact of this repository
License
See the license field above.