https://github.com/briancaffey/nvidia-nim-kit
A toolkit to explore and build with NVIDIA NIMs
https://github.com/briancaffey/nvidia-nim-kit
3d-models asr fastapi image-generation llm-inference nuxt nvidia-nim object-detection
Last synced: 9 months ago
JSON representation
A toolkit to explore and build with NVIDIA NIMs
- Host: GitHub
- URL: https://github.com/briancaffey/nvidia-nim-kit
- Owner: briancaffey
- License: apache-2.0
- Created: 2025-09-17T15:48:05.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-10-01T01:22:32.000Z (9 months ago)
- Last Synced: 2025-10-01T03:27:49.745Z (9 months ago)
- Topics: 3d-models, asr, fastapi, image-generation, llm-inference, nuxt, nvidia-nim, object-detection
- Language: Vue
- Homepage:
- Size: 13.8 MB
- Stars: 2
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
README
# NIM Kit
A toolkit to explore NVIDIA NIMs locally — from LLMs to 3D and beyond.
## Quick Start
1. **Start the backend services:**
```bash
docker compose up
```
2. **Start the frontend development server:**
```bash
cd frontend
yarn && yarn dev
```
3. **Configure NVIDIA API (optional):**
Copy `.env.sample` to `.env` and add your `NVIDIA_API_KEY` for enhanced functionality.
## Supported NIM Services
| Service Type | Model | Description |
|--------------|-------|-------------|
| **LLM** | Meta Llama 3.1 8B Instruct | Advanced language understanding and text generation |
| **LLM** | DeepSeek R1 Distill Llama 8B | Enhanced reasoning and coding capabilities |
| **Image Generation** | FLUX.1 Schnell | Fast, high-quality image generation |
| **Image Generation** | FLUX.1 Dev | State-of-the-art image generation suite |
| **Image Generation** | FLUX.1 Kontext Dev | Multimodal in-context image generation and editing |
| **3D Generation** | Microsoft Trellis | Generate 3D assets from text or images |
| **Speech Recognition** | NVIDIA Parakeet CTC 0.6B ASR | English speech-to-text transcription |
| **Speech Enhancement** | NVIDIA Studio Voice | Enhance speech quality and correct audio degradations |
| **OCR** | Baidu PaddleOCR | Optical character recognition and table extraction |
## Architecture
- **Backend**: FastAPI with Celery for async task processing
- **Frontend**: Nuxt.js with Vue 3, Tailwind CSS and shadcn-ui
- **Database**: Redis for caching and task queuing
- **Monitoring**: Flower for Celery task monitoring
## NIM Observability
Monitor your NIM deployments with built-in observability tools including Prometheus metrics, Grafana dashboards, and alerting rules. Navigate to the `nim-observability/` directory to set up comprehensive monitoring for your NIM services.