metadata
license: apache-2.0
tags:
- adam
- approx-gelu
- deit
- gated-fusion
- layernorm
- multitask
- polynomial
- sparse
- tiny
- trunc-normal
model_563478059_deit_tiny.py
Model Overview
A tiny-scale implementation of the deit architecture, built for multitask tasks.
Architecture
- Architecture: deit
- Scale: tiny
- Attention: sparse
- Fusion strategy: gated fusion
- Task head: multitask
- Activation: approx gelu
- Normalization: layernorm
- Initialization: trunc normal
Training
- Optimizer: adam
- LR scheduler: polynomial
Files
model_563478059_deit_tiny.py— main artifact of this repository
License
See the license field above.