| ---
|
| license: apache-2.0
|
| tags:
|
| - adam
|
| - approx-gelu
|
| - beit
|
| - exponential
|
| - matching
|
| - scalenorm
|
| - small
|
| - standard
|
| - trunc-normal
|
| - tucker
|
| ---
|
|
|
| # main.py
|
|
|
| ## Model Overview
|
|
|
| A **small**-scale implementation of the **beit** architecture, built for **matching** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: beit
|
| - **Scale**: small
|
| - **Attention**: standard
|
| - **Fusion strategy**: tucker
|
| - **Task head**: matching
|
| - **Activation**: approx gelu
|
| - **Normalization**: scalenorm
|
| - **Initialization**: trunc normal
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adam
|
| - **LR scheduler**: exponential
|
|
|
| ## Files
|
|
|
| - `main.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
