model_259871266_albef_giant.py

Model Overview

A giant-scale implementation of the albef architecture, built for retrieval tasks.

Architecture

• Architecture: albef
• Scale: giant
• Attention: dilated
• Fusion strategy: tucker
• Task head: retrieval
• Activation: gelu
• Normalization: batchnorm
• Initialization: kaiming

Training

• Optimizer: sgd
• LR scheduler: polynomial

Files

• model_259871266_albef_giant.py — main artifact of this repository

License

See the license field above.