metadata
license: mit
tags:
- classification
- concat-mlp
- exponential
- orthogonal
- poolformer
- relu
- rmsnorm
- sgd
- sliding-window
- small
model_420508441_poolformer_small.py
Model Overview
A small-scale implementation of the poolformer architecture, built for classification tasks.
Architecture
- Architecture: poolformer
- Scale: small
- Attention: sliding window
- Fusion strategy: concat mlp
- Task head: classification
- Activation: relu
- Normalization: rmsnorm
- Initialization: orthogonal
Training
- Optimizer: sgd
- LR scheduler: exponential
Files
model_420508441_poolformer_small.py— main artifact of this repository
License
See the license field above.