5fa1a76
1
4-bit quantization compresses a model even further, and it is commonly used with QLoRA to finetune quantized LLMs.
