| ---
|
| license: apache-2.0
|
| tags:
|
| - batchnorm
|
| - gelu-tanh
|
| - grouped-query
|
| - lion
|
| - low-rank
|
| - nano
|
| - onecycle
|
| - poolformer
|
| - retrieval
|
| - xavier-uniform
|
| ---
|
|
|
| # eval.py
|
|
|
| ## Model Overview
|
|
|
| A **nano**-scale implementation of the **poolformer** architecture, built for **retrieval** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: poolformer
|
| - **Scale**: nano
|
| - **Attention**: grouped query
|
| - **Fusion strategy**: low rank
|
| - **Task head**: retrieval
|
| - **Activation**: gelu tanh
|
| - **Normalization**: batchnorm
|
| - **Initialization**: xavier uniform
|
|
|
| ## Training
|
|
|
| - **Optimizer**: lion
|
| - **LR scheduler**: onecycle
|
|
|
| ## Files
|
|
|
| - `eval.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
