The experimental results on the homepage are based on zero-shot learning？

BlinkDL / RWKV-LM

RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.

Apache License 2.0

12.61k stars 859 forks source link

The experimental results on the homepage are based on zero-shot learning？ #52

Closed AIRedWood closed 1 year ago

AIRedWood commented 1 year ago

Dear author

I was wondering if I could ask you a question about your experiment results. Specifically, I'm curious to know if your results were based on zero-shot, one-shot, or few-shot conditions.。 experiment results lnk

BlinkDL commented 1 year ago

These are all zero-shot scores, and they are all trained on the Pile, so directly comparable