model_024934883_vit_xlarge.py

Model Overview

A xlarge-scale implementation of the vit architecture, built for generation tasks.

Architecture

• Architecture: vit
• Scale: xlarge
• Attention: sparse
• Fusion strategy: low rank
• Task head: generation
• Activation: swish
• Normalization: scalenorm
• Initialization: kaiming

Training

• Optimizer: novograd
• LR scheduler: linear warmup

Files

• model_024934883_vit_xlarge.py — main artifact of this repository

License

See the license field above.