Projects

### Overview: Transformer Lab is an open-source platform designed for AI researchers and teams to seamlessly train, evaluate, and scale machine learning models. It supports both local and cloud environments, offering a unified interface for training, fine-tuning, and evaluating models. ### Key Features: - **Unified Interface:** Seamless training, fine-tuning, and evaluation of models. - **Cross-Platform:** Runs on macOS, Linux, and Windows (WSL2). - **Local Privacy:** No data leaves your machine. - **Full Toolkit:** Includes training, fine-tuning, chat, and evaluation functionalities. - **Collaborative:** Centralized experiment tracking, model registry, and artifact management. - **Interactive Compute:** One-click Jupyter, VSCode, and SSH sessions on remote nodes. - **Resilience:** Auto-recovery from checkpoints and spot instance preemption. ### Tech Stack: Python ### Why It Matters: Transformer Lab simplifies the fragmented AI tooling landscape, providing a single, elegant interface for researchers and teams to train, evaluate, and scale models across local hardware, on-prem, or in the cloud. It supports a wide range of models and formats, making it a versatile solution for both individual researchers and research teams.

### Overview: H2O LLM Studio is a framework and no-code GUI designed for fine-tuning state-of-the-art large language models (LLMs). It simplifies the process of fine-tuning models without requiring any coding experience, providing a user-friendly interface for training and evaluating LLMs. ### Key Features: - **Graphical User Interface (GUI)**: A specialized GUI for large language models. - **Fine-Tuning Hyperparameters**: Supports a wide variety of hyperparameters for fine-tuning. - **Low-Rank Adaptation (LoRA)**: Utilizes recent fine-tuning techniques with a low memory footprint. - **8-bit Model Training**: Efficiently trains models with a low memory footprint. - **Reinforcement Learning (RL)**: Experimental feature to fine-tune models using reinforcement learning. - **Advanced Evaluation Metrics**: Measures generated answers to judge model performance. - **Model Export**: Easily export your model to the Hugging Face Hub for community sharing. - **Chat Interface**: Instant feedback on model performance. - **Data Format and Example Data**: Supports various data formats and provides example data for demonstration. ### Tech Stack: Python ### Why It Matters: H2O LLM Studio simplifies the process of fine-tuning LLMs, making it accessible to a broader audience without requiring advanced coding skills. The GUI and advanced features like LoRA and RLHF optimization make it a powerful tool for researchers and developers looking to fine-tune their models effectively.

### Overview: Argilla is a collaboration tool designed for AI engineers and domain experts to build high-quality datasets for their projects. ### Key Features: - **Collaborative Data Collection**: Teams can work together to collect and label data from multiple sources. - **Automated Feedback**: AI-driven suggestions for improving data quality and relevance. - **Semantic Search**: Easily find and label data based on context and intent. - **Workflow Integration**: Seamless integration with existing workflows and tools. ### Tech Stack: Python ### Why It Matters: Argilla helps ensure high-quality datasets, improving the output quality of AI models and empowering teams to manage their data and models efficiently.

### Overview: The Bard API project is an unofficial Python wrapper that allows developers to interact with Google's Bard through the use of a cookie value. This project aims to provide a simple and lightweight solution for developers who face frequent authentication issues or are unable to authenticate using Google Authentication. ### Key Features: - **Cookie-Based Interaction**: Utilizes the cookie value to interact with Google Gemini, bypassing the need for frequent authentication. - **Reverse Engineering**: Based on the inference process of Google Bard, enabling developers to ask questions and receive responses. - **Minimal Dependencies**: Requires only Python to be installed, making it easy to integrate into existing projects. - **Community Collaboration**: Collaborated with Antonio Cheong to enhance the project and improve its functionality. ### Tech Stack: - Python ### Why It Matters: This project is crucial for developers who need to integrate Google Bard functionality into their applications without encountering authentication issues. It simplifies the process of using Google's large language model for various applications, making it accessible to a broader audience. ### Repository: The repository is maintained at [dsdanielpark/Bard-API](https://github.com/dsdanielpark/Bard-API). ### Installation: ```bash pip install python-gemini-api ``` ```bash pip install git+https://github.com/dsdanielpark/Gemini-API.git ``` For the updated version, use: ```bash pip install -q -U python-gemini-api ``` ### Notable Collaboration: The project has been collaboratively developed with Antonio Cheong, enhancing