metadata
license: mit
tags:
- blip
- contrastive
- dilated
- gated-fusion
- gelu
- lamb
- polynomial
- rmsnorm
- tiny
- xavier-uniform
model_157588127_blip_tiny.py
Model Overview
A tiny-scale implementation of the blip architecture, built for contrastive tasks.
Architecture
- Architecture: blip
- Scale: tiny
- Attention: dilated
- Fusion strategy: gated fusion
- Task head: contrastive
- Activation: gelu
- Normalization: rmsnorm
- Initialization: xavier uniform
Training
- Optimizer: lamb
- LR scheduler: polynomial
Files
model_157588127_blip_tiny.py— main artifact of this repository
License
See the license field above.