A collection of quantized models. All the models can be fine-tuned by adding a LoRA Adapter.