| ---
|
| license: cc-by-4.0
|
| tags:
|
| - adafactor
|
| - bilinear
|
| - contrastive
|
| - gelu
|
| - grouped-query
|
| - kaiming
|
| - linear-warmup
|
| - mocov3
|
| - rmsnorm
|
| - small
|
| ---
|
|
|
| # model_719180533_mocov3_small.py
|
|
|
| ## Model Overview
|
|
|
| A **small**-scale implementation of the **mocov3** architecture, built for **contrastive** tasks.
|
|
|
| ## Architecture
|
|
|
| - **Architecture**: mocov3
|
| - **Scale**: small
|
| - **Attention**: grouped query
|
| - **Fusion strategy**: bilinear
|
| - **Task head**: contrastive
|
| - **Activation**: gelu
|
| - **Normalization**: rmsnorm
|
| - **Initialization**: kaiming
|
|
|
| ## Training
|
|
|
| - **Optimizer**: adafactor
|
| - **LR scheduler**: linear warmup
|
|
|
| ## Files
|
|
|
| - `model_719180533_mocov3_small.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
