model_647883246_mae_base.py
Model Overview
A base-scale implementation of the mae architecture, built for multitask tasks.
Architecture
- Architecture: mae
- Scale: base
- Attention: flash
- Fusion strategy: tensor fusion
- Task head: multitask
- Activation: gelu tanh
- Normalization: scalenorm
- Initialization: orthogonal
Training
- Optimizer: lamb
- LR scheduler: exponential
Files
model_647883246_mae_base.py — main artifact of this repository
License
See the license field above.
