Modelpublic

HimanshuPewal24/bug-explainer-mistral_qlora

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face

0likes2downloads

Model Card

🐛 Bug Explainer — Mistral-7B QLoRA

Finetuned from mistralai/Mistral-7B-Instruct-v0.2 using 4-bit QLoRA on Stack Overflow error/answer pairs. Supports Python · JavaScript · Java · C++.

Training details

—Quantization: 4-bit NF4 + double quantization
—LoRA rank: 16, alpha: 32
—Epochs: 3
—Optimizer: pagedadamw32bit