model_370332861_mobilevit_giant.py

Model Overview

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

Architecture

• Architecture: mobilevit
• Scale: giant
• Attention: multi query
• Fusion strategy: tensor fusion
• Task head: retrieval
• Activation: gelu
• Normalization: layernorm
• Initialization: kaiming normal

Training

• Optimizer: novograd
• LR scheduler: exponential

Files

• model_370332861_mobilevit_giant.py — main artifact of this repository

License

See the license field above.