This is a Hugging Face transformers-style conversion of the original SMoE 15B-parameter model with BFLOAT16 from the paper "Efficient Large Scale Language Modeling with Mixtures of Experts" from Artetxe et al. The original model card can be found at https://github.com/facebookresearch/fairseq/blob/main/examples/moe_lm/model_card.md.
The usage example and modeling code can be found at https://github.com/pingzhili/light-fairseq
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.