sonicTheDatahog/Llama-2-7b-chat-finetune-health-chatbot
2
This is a finetuned model I have made using QLora (PEFT) with lora_r = 64 (help decide the matrice size where the weights will change) and lora_alpha = 16 (defines the intensity with which the lora will change the paramters) which are the majorly important hyperparameters of QLora which I learnt recently with the help of Krish Naik.