Very impressive. WizardCoder 34B outperforms GPT-4, Claude, and Bard on HumanEval. It also surpasses all open-source Code LLMs by a substantial margin. WizardCoder 34B is fine-tuned version of Code Llama model using Evol-Instruct, a synthetic dataset. The code, model weights,… https://t.co/LRio9LQIkz https://t.co/kq7mpenrv1
— Lior⚡ (@AlphaSignalAI) Aug 26, 2023
from Twitter https://twitter.com/AlphaSignalAI
August 26, 2023 at 10:51AM
via IFTTT
No comments:
Post a Comment