| ---
|
| license: mit
|
| tags:
|
| - co-attention
|
| - efficientformer
|
| - exponential
|
| - instancenorm
|
| - kaiming-normal
|
| - lamb
|
| - linear
|
| - mish
|
| - retrieval
|
| - tiny
|
| ---
|
|
|
| # main.py
|
|
|
| ## Model Overview
|
|
|
| A **tiny**-scale implementation of the **efficientformer** architecture, built for **retrieval** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: efficientformer
|
| - **Scale**: tiny
|
| - **Attention**: linear
|
| - **Fusion strategy**: co attention
|
| - **Task head**: retrieval
|
| - **Activation**: mish
|
| - **Normalization**: instancenorm
|
| - **Initialization**: kaiming normal
|
|
|
| ## Training
|
|
|
| - **Optimizer**: lamb
|
| - **LR scheduler**: exponential
|
|
|
| ## Files
|
|
|
| - `main.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
