Incotelligence SimChat

This repository contains the code for Incotelligence SimChat, an experimental chatbot trained on a small custom dataset (data/dialogues.txt) and is not yet intended for production use. This version is meant for testing, research and learning purposes and most responses are not perform well in all scenarios like handling ambiguous queries or providing detailed explanations.

Running the Chatbot

Make sure you are using Python 3.13 or higher

If you have a CUDA-compatible GPU, PyTorch will use it automatically. Otherwise, the chatbot will run on CPU (slower).

1. Clone this repository
2. Install the required dependencies:

pip install -r requirements.txt

3. Train the model:

If you already have the pretrained model model.pt, place it in src/models/ and skip step 3.

python src/train.py

4. Run the chatbot:

python src/main.py

5. Interact with the chatbot via the command line.

Troubleshooting

• If you get a ModuleNotFoundError, ensure all dependencies from requirements.txt are installed.
• On Windows, you might need to install torch with the proper CUDA version manually.