Finetune bigScience-bloom-1B7 model with Persian News Summary

You can check all the steps in the notebook or you can run the notebook on kaggle if you want to test it out

• Because this model is big enough to make most of GPUs out-of-memory, I used LoRA and 4-bit Quantization