MRzzm/DINet

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1128 stars · 181 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

DINet is the official PyTorch implementation of an AAAI 2023 paper on realistic face visually dubbing, which deforms and inpaints mouth regions in high-resolution video to match driving audio. It includes inference scripts with pretrained models and a coarse-to-fine training pipeline for arbitrary resolutions.

Use cases

• dub a talking-head video to match new audio
• generate lip-synced face video from speech
• replace mouth region in high resolution video driven by audio
• train an audio-driven face dubbing model on custom dataset
• process HDTF dataset videos for talking face generation

When to choose

• you need research-grade audio-driven face dubbing with high resolution output
• you want to retrain or fine-tune a talking-face model on your own audio-visual data
• your videos are frontal-view talking heads with normal lighting

When to avoid

• you need a production-ready tool with strong generalization to arbitrary faces and angles
• you need a no-GPU or plug-and-play consumer application
• you require a maintained project with a license or active development

Facets

library · maturity maintenance

deep-learning video-processing image-processing audio-processing machine-learning computer-vision deep-learning artificial-intelligence python windows talking-face face-dubbing lip-sync video-inpainting research-code aaai2023 audio-driven video audio linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="MRzzm/DINet")