imoneoi/openchat
OpenChat provides open-source large language models fine-tuned with C-RLFT, a reinforcement learning strategy for learning from mixed-quality data.
Velocity · 7d
+4.9
★ / day
Trend
→steady
star history
OpenChat develops and releases open-source language models, including 7B-parameter variants that run on consumer GPUs. The models are fine-tuned using C-RLFT, an offline reinforcement learning approach that learns from mixed-quality data without requiring preference labels. Multiple model versions have been released, including recent Llama-3 based variants that outperform official Llama 3 instruct models and other open-source fine-tunes.