https://github.com/sofialtamsh/nlp-chatbot

NLP-Chatbot
https://github.com/sofialtamsh/nlp-chatbot

chatbot cuda generative-network machine-learning natural-language-processing nlp nlp-python nlp-question-answering python question-answering question2answer seq2seq squad-dataset tensorflow

Last synced: 12 days ago
JSON representation

NLP-Chatbot

• Host: GitHub
• URL: https://github.com/sofialtamsh/nlp-chatbot
• Owner: sofialtamsh
• Created: 2024-11-15T09:39:03.000Z (3 months ago)
• Default Branch: main
• Last Pushed: 2024-11-30T17:35:00.000Z (2 months ago)
• Last Synced: 2024-11-30T18:30:12.232Z (2 months ago)
• Topics: chatbot, cuda, generative-network, machine-learning, natural-language-processing, nlp, nlp-python, nlp-question-answering, python, question-answering, question2answer, seq2seq, squad-dataset, tensorflow
• Language: Python
• Homepage:
• Size: 7.81 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Question Answering On SQUAD

## Short-Video

https://user-images.githubusercontent.com/87471617/140063943-754d35aa-e806-4234-a5b4-108b2193d962.mp4

## Requirements

- Tensorflow=2.4.0

- CUDA=11.2

- CUDNN=11.4 (necessary only for gpu computing (optional))

- Python 3.6

## Dataset

The Stanford Question Answering Dataset (SQuAD), which is derived from Wikipedia, can be used for question answering chatbot. The SQuAD includes:

- 107,785 question-answer pairs depend on 536 articles.

- Due to a lack of RAM, only 10.000 pairs have been used for training of the Seq2Seq model.

- According to the results, given questions to the model can be predicted by the model accurately.

- Given data to the model should be enhanced in order to increase the accuracy of the model.

## Folders:

### Data: 

- Downloaded data will be saved in this folder. Dowloading script is available in "Preporeccsing file.py". The program will download it in Data folder, if the folder exists.

### Weight:

- Model weight, encoder and decoder model will be saved in this folder.

### Image:

- The history of the model will be saved in this folder.

## Model:

Seq2Seq architecture has been chosen to train the model.

## Running:

- The "main.py" file must be run in a virtual environment with the system requirements.