Training procedure

This adapter has been fine-tuned using quantization-aware LoRA (QA-LoRA).

More details on the training procedure here: Fine-tune Quantized Llama 2 on Your GPU with QA-LoRA

The base model was quantized with AutoGPTQ INT4. You can find it here:

kaitchup/Llama-2-7b-4bit-32g-autogptq

Framework versions

• PEFT 0.4.0