AI Agent Loader

A local AI model runner with a browser-based GUI. Browse, download, and run Hugging Face models on your own hardware.

Features

• Hub Browser — Search and explore Hugging Face models with filters for task type, library, and popularity. View model cards, file sizes, and download with one click.
• Text Generation — Run LLMs (Llama, Mistral, Phi, etc.) with chat and completion modes. Supports both transformers and quantized GGUF models.
• Image Generation — Generate images with Stable Diffusion, SDXL, Flux, and other diffusion models.
• Classification — Text classification, sentiment analysis, image classification, and named entity recognition.
• Speech — Speech-to-text (Whisper) and text-to-speech.
• Object Detection — Detect objects in images with bounding boxes.
• Summarization & Translation — Summarize long texts and translate between languages.

Requirements

• Python 3.10+
• NVIDIA GPU with CUDA support (recommended)
• ~2GB disk space for the app + space for downloaded models

Setup

1. Clone the repository:

   git clone <repo-url>
   cd AI_agent_loader
   

2. Create a virtual environment:

   python -m venv venv
   venv\Scripts\activate  # Windows
   # source venv/bin/activate  # Linux/Mac
   

3. Install PyTorch with CUDA:

   pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
   

4. Install dependencies:

   pip install -r requirements.txt
   

5. (Optional) GGUF support with GPU acceleration:

   set CMAKE_ARGS=-DGGML_CUDA=on
   pip install llama-cpp-python --force-reinstall --no-cache-dir
   

6. (Optional) Set up Hugging Face token for gated models:

   copy .env.example .env
   # Edit .env and add your HF token
   

Usage

python app.py

Open your browser to http://localhost:7860.

Quick Start

1. Go to the Hub Browser tab
2. Search for a model (e.g., "TinyLlama" for text generation)
3. Click a result to see details, then click Download Model
4. Switch to the appropriate tab (e.g., Text Generation)
5. Click Refresh, select the model, click Load
6. Enter a prompt and click Generate

Architecture

The app uses a modular runner system. Each AI task type has its own runner that implements a common interface (load, run, unload). Adding support for a new model type requires:

1. Create a new runner in runners/
2. Create a UI tab in ui/
3. Register the pipeline tag in config/task_registry.py
4. Add the tab to ui/app_builder.py