Overview

This repository contains two Jupyter Notebook files that demonstrate the use of the Zephyr-7b-alpha model and the neuralChat model from Hugging Face for text classification tasks.

Files in this Repository

• Example_FirstStep_zephyr-7b-alpha.ipynb: A beginner-friendly guide to getting started with the Zephyr-7b-alpha model.
• Example20231211_tweets_sentiment_neuralChat.ipynb: An advanced example showcasing more complex uses of the model, including sentiment analysis on tweets, optimized for Google Colab's A100 GPU.

Example_FirstStep_zephyr-7b-alpha.ipynb

Steps Covered

• Installation: Instructions to install necessary libraries (transformers and accelerate) from Hugging Face.
• Load Dependencies: Importing torch and the pipeline from transformers.
• Model Loading: Initializing the text generation pipeline with the Zephyr-7b-alpha model.
• First Prompt Setup: Demonstrates how to set up a prompt for text generation, including defining roles (system, user, assistant) for the interaction.
• Response Generation: Shows how to generate a response from the model based on the provided prompt.
• Functions for Text Classification: Includes functions for zero-shot and one-shot text classification, demonstrating the versatility of the model.
• Example Runs: Provides example runs for both zero-shot and few-shot text classification using the model.

Example20231211_tweets_sentiment_neuralChat.ipynb

This notebook delves deeper into using the neuralChat model for sentiment analysis of tweets. It is optimized for Google Colab's A100 GPU.

• Link: https://docs.google.com/presentation/d/1jCLwH9vInX6RDy00t7HISQPvLXkPebms-qu_ZyCrqDM/edit?usp=sharing