I did an educational purpose small attempt to use quantization method to fine tune on QA dataset using Colab provided GPU with LLama2-7B. Although Quantizing model is not a challenge but fine tuning is 😅. I would love to hear some criticism and talks about it and if you ever tried it I also have some questions 😇 https://github.com/UmarIgan/Machine-Learning/blob/master/Quantize_Llama2_7B_model_and_Fine_Tune_on_QA_Datasets.ipynb