| ---
|
| license: cc-by-4.0
|
| tags:
|
| - adafactor
|
| - batchnorm
|
| - concat-mlp
|
| - contrastive
|
| - dilated
|
| - exponential
|
| - gelu-tanh
|
| - mae
|
| - xavier
|
| - xlarge
|
| ---
|
|
|
| # model.py
|
|
|
| ## Model Overview
|
|
|
| A **xlarge**-scale implementation of the **mae** architecture, built for **contrastive** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: mae
|
| - **Scale**: xlarge
|
| - **Attention**: dilated
|
| - **Fusion strategy**: concat mlp
|
| - **Task head**: contrastive
|
| - **Activation**: gelu tanh
|
| - **Normalization**: batchnorm
|
| - **Initialization**: xavier
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adafactor
|
| - **LR scheduler**: exponential
|
|
|
| ## Files
|
|
|
| - `model.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
