Welcome to the Barebone-Boogie-Chatbot project!
This project utilizes the power of the transformers and BLOOM-style models to create a local chatbot.
The chatbot is designed as a developmental tool for machine learning enthusiasts and researchers.
For a polished comprehensive local chatbot solution, consider Freedom GPT instead.
Barebone-Boogie-Chatbot is an efficient compact command-line chatbot designed to evaluate and test different machine learning models.
The Barebone-Boogie-Chatbot is a Python-based chatbot project that uses the transformers library and the AutoModelForCausalLM model. It has the capability to run a chatbot with a Hugging Face model or a local custom model.
This project serves as a resourceful tool to explore the possibilities of AI in the field of Natural Language Processing (NLP).
This project is designed in such a way that machine learning enthusiasts and researchers can tweak and adjust the configurations to effectively assess their new models.
This minimal local chatbot solution can serve as a foundation for more complex projects.
- Model Flexibility: The software allows users to input different pre-trained models by name. This can be a local model or models from the Hugging Face library.
- Adjustable Parameters: It offers control over several parameters such as temperature, top_k, maximum response length, and the number of return sequences to enhance the precision and customization of the model’s outputs.
- Hardware Compatibility: While it's designed to run optimally on a GPU for faster computation, it also works with CPUs, ensuring functionality across a variety of hardware setups.
Software Requirements: This software requires Python 3.8 or newer, as the minimum version of the transformers library we have implemented (version 4.25.0) is not compatible with earlier Python versions.
To get a local copy up and running, follow these simple steps:
- Clone the repo
git clone https://github.com/mikkorautiainen/Barebone-Boogie-Chatbot.git- Install the required packages
pip3 install -r requirements.txtThe chatbot is composed of two main Python scripts:
barebone_chatbot.py: This is the main script that handles user interaction and utilizes the model to generate responses to user's input.barebone_response_ranker.py: This script contains a class ResponseRanker that ranks the generated responses based on their suitability to the user's question.
To start using the Barebone-Boogie-Chatbot, navigate to the project directory and run the following command:
python3 barebone_chatbot.py-h, --help: Displays the help message and exits the program.--model_name MODEL_NAME: Specifies the pre-trained model to use.--temperature TEMPERATURE: Controls the randomness of the model’s outputs.--top_k TOP_K: Determines the number of most likely next words the model samples from.--max_length MAX_LENGTH: Sets the maximum length of the generated responses.--num_return_sequences NUM_RETURN_SEQUENCES: Defines the number of sequences to return.--debug: Enable debug mode.
The chatbot is highly customizable. You can adjust parameters like temperature, the number of top K most likely next words to sample from, the maximum length of the generated sequences and the number of sequences to return.
For instance, to run the chatbot using the TurkuNLP/gpt3-finnish-xl model with a temperature of 0.7, top_k of 500, and a maximum response length of 50, the command would be:
python3 barebone_chatbot.py --model_name TurkuNLP/gpt3-finnish-xl --temperature 0.7 --top_k 500 --max_length 50To exit the chatbot, enter 'lopeta' as your input.
Contributions are what make the open-source community such an amazing place to learn, inspire, and create.
Any contributions you make are greatly appreciated !!!
- Fork the Project
- Create your Feature Branch (git checkout -b feature/AmazingFeature)
- Commit your Changes (git commit -m 'Add some AmazingFeature')
- Push to the Branch (git push origin feature/AmazingFeature)
- Open a Pull Request
Distributed under the MIT License. See LICENSE for more information.