| ---
|
| license: cc-by-4.0
|
| tags:
|
| - adafactor
|
| - generation
|
| - low-rank
|
| - nano
|
| - scalenorm
|
| - standard
|
| - step
|
| - swish
|
| - vit
|
| - xavier
|
| ---
|
|
|
| # pipeline.py
|
|
|
| ## Model Overview
|
|
|
| A **nano**-scale implementation of the **vit** architecture, built for **generation** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: vit
|
| - **Scale**: nano
|
| - **Attention**: standard
|
| - **Fusion strategy**: low rank
|
| - **Task head**: generation
|
| - **Activation**: swish
|
| - **Normalization**: scalenorm
|
| - **Initialization**: xavier
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adafactor
|
| - **LR scheduler**: step
|
|
|
| ## Files
|
|
|
| - `pipeline.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
