TorchUtility: A Utility Toolkit for PyTorch, as the name suggests, this is a toolset that helps make better use of the pyTorch framework, it aims to be a comprehensive collection of tools and utilities designed to complement the pyTorch framework.
Under development, not yet released!
-
Utility: A Comprehensive Toolkit for Developers, includes Advanced Data Operations,... etc.
-
Security: Unique utilities for encrypting data and models, ensuring secure model training and deployment in sensitive applications.
-
llm_lighting: A streamlined module dedicated to simplifying the complexities of Large Language Model (LLM) integration, focusing on efficient training, fine-tuning, and inference with minimal code. This component utilizes a one-file approach for core functionalities, enhancing the project's utility with lightweight processes and seamless interaction with leading open-source libraries like Hugging Face. causallm_lora_train.py: causallm_lora_inference.py: This script offers a straightforward and minimalistic approach to performing inference with LoRA-optimized Causal Language Models, ensuring high efficiency with lean code.
Usage Disclaimer: This project is for educational and research purposes only. You are responsible for any risks associated with using this project. The author shall not be liable for any direct or indirect losses caused by the use of this project. Please make sure you understand and agree to these terms before using this project.
Contribution Disclaimer: The author welcomes any form of contribution, but shall not be responsible for any direct or indirect losses caused by submitting contributions. Please ensure you understand and agree to these terms before submitting any contributions. Contributors must ensure that their contributions comply with relevant laws and regulations and respect the intellectual property rights of others.
This project is licensed under the BSD 3-clause License. See the LICENSE file for more details.
Contain codes written with the assistance of LLM (such as GPT) assistants