| ---
|
| license: cc-by-4.0
|
| tags:
|
| - adafactor
|
| - batchnorm
|
| - bilinear
|
| - contrastive
|
| - mish
|
| - mobilevit
|
| - onecycle
|
| - sliding-window
|
| - small
|
| - xavier-uniform
|
| ---
|
|
|
| # train.py
|
|
|
| ## Model Overview
|
|
|
| A **small**-scale implementation of the **mobilevit** architecture, built for **contrastive** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: mobilevit
|
| - **Scale**: small
|
| - **Attention**: sliding window
|
| - **Fusion strategy**: bilinear
|
| - **Task head**: contrastive
|
| - **Activation**: mish
|
| - **Normalization**: batchnorm
|
| - **Initialization**: xavier uniform
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adafactor
|
| - **LR scheduler**: onecycle
|
|
|
| ## Files
|
|
|
| - `train.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
