Guanaco vs. Pythia

Guanaco

Guanaco is an LLM based off the QLoRA 4-bit finetuning method developed by Tim Dettmers et. al. in the UW NLP group. Guanaco achieves 99% ChatGPT performance on the Vicuna benchmark.

Guanaco is an LLM that uses a finetuning method called LoRA that was developed by Tim Dettmers et. al. in the UW NLP group. With QLoRA, it becomes possible to finetune up to a 65B parameter model on a 48GB GPU without loss of performance relative to a 16-bit model. The Guanaco model family outperforms all previously released models on the Vicuna benchmark. However, given the models are based off of the LLaMA model family, commercial use is not permitted.

Further reading

• arxiv.org/abs/2305.14314
• twitter.com/Tim_Dettmers/status/1661379354507476994

Pythia

The most recent (as of May 2023) effort from EleutherAI, Pythia is a set of LLMs trained on The Pile. While it appears to outperform OPT and GPTNeo, its performance against GPT-J is unclear. Versions of Pythia have also been instruct-tuned by the team at Together.