Message_Chatbot

Purpose

This script demonstrates creating a model using a data set of question-answer pairs extracted from individual Facebook conversations in Vietnamese that are preprocessed. from Hugging Face and use it to automatically reply to messages on Facebook.

Function

• The data set is created from personal Facebook chat logs in Vietnamese.
• Model creation is pre-processed with model data that has been pre-trained on a data set of question-answer pairs.
• Use selenium to automatically reply to conversations

Analysis

1. Extract data:

• Download personal information from Facebook account.
• Use Python to extract and pre-process conversations from downloaded Facebook data to create a refined dataset.
• Preprocessing data

2. Training:

• Script to customize the model:
• Complete dataset over 10 epochs.

Communicate

• CPU device:
• Please note that training on CPU may result in long waiting times due to the computational intensity of model fine-tuning.

Use

1. Prepare data:

• Make sure your dataset of question-answer pairs is saved in fb_messages.txt.

2. Run fine-tuning:

• Execute the script to start the adjustment process.
• The code includes: data creation, training, use, and automatic application with elenium

3. Create feedback:

• Use the refined model to generate answers to new questions using the code provided.