Zephyr
Zephyr is a family of open, instruction-tuned chat models developed by Hugging Face's H4 team, built by fine-tuning Mistral 7B using Direct Preference Optimization (DPO) on distilled AI-generated preference data, and notable for strong chat performance relative to its small parameter size.
7 resources across 1 library
Glossary Terms(7)
WizardLM
WizardLM is a family of open instruction-tuned language models notable for introducing 'Evol-Instruct,' a method that uses a large language model to automatica…
OpenChat
OpenChat is an open-source language model project that fine-tunes base models such as Mistral 7B using C-RLFT (Conditioned Reinforcement Learning Fine-Tuning),…
Zephyr
Zephyr is a family of open, instruction-tuned chat models developed by Hugging Face's H4 team, built by fine-tuning Mistral 7B using Direct Preference Optimiza…
Orca (model)
Orca is a series of research language models from Microsoft that use explanation-tuning — training on step-by-step reasoning traces distilled from a larger tea…
Zephyr (model)
Zephyr is a series of open large language models from Hugging Face fine-tuned from Mistral 7B using distilled supervised fine-tuning and Direct Preference Opti…
StableLM
StableLM is a family of open-source language models developed by Stability AI, released in multiple sizes and generations to provide permissively licensed, tra…
Dolly (model)
Dolly is an open-source instruction-following large language model released by Databricks, notable as one of the first LLMs whose instruction-tuning dataset wa…