fine-tuning the NousResearch/llama-2-7b-chat-hf model on a custom dataset using QLoRA (Quantization in Language Representation) Peft (Position Embedding Fourier Transform) techniques. (Supervised Fine-tuning) approach is employed for training.
Python 3.9 PyTorch Hugging Face Transformers Datasets Peft Trl
Install Dependencies: pip install -r requirements.txt
Download LLMA2 Pre-trained Model: Obtain the LLMA2 pre-trained model ("NousResearch/llama-2-7b-chat-hf") from the Hugging Face model hub.
The LLMA2 model and its components, QLoRA and Peft, were developed by NousResearch. Please refer to the official documentation for more details.
- run this model on t4 gpu on colab for faster & smoother running keep describe version which want to use bocz
2)if u install like this !pip install torch it install but when u run training arguments they show you you not install torch / accelerate thats why you mention version of that in my case i use torch==1.10.0 like this not this !pip install torch