HunyuanVideo-I2V

HunyuanVideo-I2V, from Tencent, is a open-source image-to-video generation model. Up to 720p resolution, 129 frames. Supports custom LoRA training for unique effects.

HunyuanVideo-I2V by Tencent is an open-source image-to-video generation model capable of producing high-resolution videos up to 720p and 129 frames (5 seconds). Leveraging LoRA (Low-Rank Adaptation) technology, it allows users to train custom effects for unique video outputs. The framework includes pre-trained models, inference code, and LoRA training scripts, enabling seamless integration of image and text inputs for coherent video generation. Designed for advanced video creation, it supports parallel inference on multiple GPUs for faster processing. Ideal for developers and researchers, HunyuanVideo-I2V fosters innovation in the open-source community by offering tools for customizable, high-quality video synthesis.