Show HN: Lance – image/video generation and understanding in one model
Summary
Show HN highlights Lance, a 3B-parameter unified multimodal model from ByteDance that handles image and video understanding, generation, and editing in a single framework. The project is trained from scratch within a 128-A100-GPU budget, includes a range of demos (text-to-video, video editing, image understanding), and provides downloadable weights and a unified inference pipeline. Benchmarks and a full developer workflow are shared, emphasizing accessibility for research and practical use on consumer hardware with detailed setup instructions.