Search Results for "deep"

HunyuanVideo is a cutting-edge framework designed for large-scale video generation, leveraging advanced AI techniques to synthesize videos from various inputs. It is implemented in PyTorch, providing pre-trained model weights and inference code for efficient deployment. The framework aims to push the boundaries of video generation quality, incorporating multiple innovative approaches to improve the realism and coherence of the generated content. Release of FP8 model weights to reduce GPU...

Vidi is a family of large multimodal models developed for deep video understanding and editing tasks, integrating vision, audio, and language to allow sophisticated querying and manipulation of video content. It’s designed to process long-form, real-world videos and answer complex queries such as “when in this clip does X happen?” or “where in the frame is object Y during that moment?” — offering temporal retrieval, spatio-temporal grounding (i.e. locating objects over time + space), and even video question answering. ...

Implementation of Make-A-Video, new SOTA text to video generator from Meta AI, in Pytorch. They combine pseudo-3d convolutions (axial convolutions) and temporal attention and show much better temporal fusion. The pseudo-3d convolutions isn't a new concept. It has been explored before in other contexts, say for protein contact prediction as "dimensional hybrid residual networks". The gist of the paper comes down to, take a SOTA text-to-image model (here they use DALL-E2, but the same learning...

...DeepFaceLab is an open-source deepfake system that enables users to swap the faces on images and on video. It offers an imperative and easy-to-use pipeline that even those without a comprehensive understanding of the deep learning framework or model implementation can use; and yet also provides a flexible and loose coupling structure for those who want to strengthen their own pipeline with other features without having to write complicated boilerplate code. DeepFaceLab can achieve results with high fidelity that are indiscernible by mainstream forgery detection approaches. ...