metadata
license: bsd-3-clause
tags:
- exponential
- linear
- novograd
- poolformer
- retrieval
- rmsnorm
- swish
- tensor-fusion
- xavier
- xlarge
pipeline.py
Model Overview
A xlarge-scale implementation of the poolformer architecture, built for retrieval tasks.
Architecture
- Architecture: poolformer
- Scale: xlarge
- Attention: linear
- Fusion strategy: tensor fusion
- Task head: retrieval
- Activation: swish
- Normalization: rmsnorm
- Initialization: xavier
Training
- Optimizer: novograd
- LR scheduler: exponential
Files
pipeline.py— main artifact of this repository
License
See the license field above.