| ---
|
| license: cc-by-4.0
|
| tags:
|
| - adam
|
| - generation
|
| - grouped-query
|
| - low-rank
|
| - mish
|
| - orthogonal
|
| - poolformer
|
| - rmsnorm
|
| - step
|
| - xlarge
|
| ---
|
|
|
| # model.py
|
|
|
| ## Model Overview
|
|
|
| A **xlarge**-scale implementation of the **poolformer** architecture, built for **generation** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: poolformer
|
| - **Scale**: xlarge
|
| - **Attention**: grouped query
|
| - **Fusion strategy**: low rank
|
| - **Task head**: generation
|
| - **Activation**: mish
|
| - **Normalization**: rmsnorm
|
| - **Initialization**: orthogonal
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adam
|
| - **LR scheduler**: step
|
|
|
| ## Files
|
|
|
| - `model.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
