model_591431181_deit_giant.py
Model Overview
A giant-scale implementation of the deit architecture, built for classification tasks.
Architecture
- Architecture: deit
- Scale: giant
- Attention: sparse
- Fusion strategy: bilinear
- Task head: classification
- Activation: gelu tanh
- Normalization: scalenorm
- Initialization: orthogonal
Training
- Optimizer: adamw
- LR scheduler: constant warmup
Files
model_591431181_deit_giant.py — main artifact of this repository
License
See the license field above.
