model_170401507_blip_large.py

Model Overview

A large-scale implementation of the blip architecture, built for multitask tasks.

Architecture

• Architecture: blip
• Scale: large
• Attention: multi query
• Fusion strategy: gated fusion
• Task head: multitask
• Activation: swish
• Normalization: batchnorm
• Initialization: orthogonal

Training

• Optimizer: adamw
• LR scheduler: exponential

Files

• model_170401507_blip_large.py — main artifact of this repository

License

See the license field above.