model_584474895_deit_xlarge.py

Model Overview

A xlarge-scale implementation of the deit architecture, built for matching tasks.

Architecture

• Architecture: deit
• Scale: xlarge
• Attention: grouped query
• Fusion strategy: bilinear
• Task head: matching
• Activation: gelu tanh
• Normalization: groupnorm
• Initialization: trunc normal

Training

• Optimizer: lamb
• LR scheduler: polynomial

Files

• model_584474895_deit_xlarge.py — main artifact of this repository

License

See the license field above.