NVIDIA Jetson AI Labprovides various tools and teaching models to help developers implement AI applications on the Jetson platform. It combines powerful Jetson hardware, advanced AI models and development resources and ecosystems to achieve innovation in automation systems and edge computing. Whether it is developing next-generation robots or real-time video analysis, the Jetson platform provides the ability to implement these AI solutions. Tools and solutions required for the project.

NVIDIA Jetson AI Lab provides optimized tools and comprehensive teaching models including:

• Text Generation (LLM): Explore the application of deep learning language models pre-trained with large amounts of data, such as text generation and language models. 

• Text + Vision (VLM)：A multi-modal model tutorial that combines visual and language understanding, including the application of real-time streaming media. 

• Image Generation ：Use models such as Stable Diffusion to generate high-quality images. 

• Visual Transformer (ViT)：An efficient visual transformer covering high-resolution dense prediction. 

• Vector database: An example of a vector database that handles multimodal data. 

• Whisper: Covers automatic speech recognition and high-quality audio generation models, such as Whisper and AudioCraft. 

• Training and resources: Developers can obtain a large amount of training materials, documents and teaching models. 

》Highlight – Multi-modal voice chat with Jetson Orin

Use Riva ASR/TTS to interactively talk about images using the latest LLaVA-1.5 13B visual language model on Jetson AGX Orin.

NVIDIA Jetson AI Lab Cases Study

• Use ASR, TTS, LLM, Audio2Face and NVIDIA Omniverse to generate AI talking avatars 

• Use text prompts for object detection and segmentation of any content on Jetson Orin

More about projects please refer: NVIDIA Jetson AI Lab (jetson-ai-lab.com)