model_481991774_swin_t_giant.py
Model Overview
A giant-scale implementation of the swin t architecture, built for matching tasks.
Architecture
- Architecture: swin t
- Scale: giant
- Attention: grouped query
- Fusion strategy: co attention
- Task head: matching
- Activation: relu
- Normalization: rmsnorm
- Initialization: xavier
Training
- Optimizer: rmsprop
- LR scheduler: constant warmup
Files
model_481991774_swin_t_giant.py — main artifact of this repository
License
See the license field above.
