metadata
license: mit
tags:
- adafactor
- clip
- cosine
- giant
- layernorm
- low-rank
- mish
- retrieval
- sliding-window
- trunc-normal
eval.py
Model Overview
A giant-scale implementation of the clip architecture, built for retrieval tasks.
Architecture
- Architecture: clip
- Scale: giant
- Attention: sliding window
- Fusion strategy: low rank
- Task head: retrieval
- Activation: mish
- Normalization: layernorm
- Initialization: trunc normal
Training
- Optimizer: adafactor
- LR scheduler: cosine
Files
eval.py— main artifact of this repository
License
See the license field above.