Lip-Wise v1.2

⬆️Significant Update for Facial Restoration!!!

The thing that's been nagging at me since this project kicked off was the subpar quality I was getting from GFPGAN and other face restoration tools. I've been consistently working with 512x512 faces, but after streamlining my process and running numerous tests, I've finally nailed down the ideal input size for these models.

Do let me know if you find this update helpful...

Full Changelog: v1.0...v1.2

Lip-Wise v1.0

I'm excited to announce the first stable release of Lip-Wise, the culmination of my efforts to create a powerful and user-friendly lip-syncing solution. This version represents the most polished iteration yet, laying the groundwork for exciting advancements in the future.

Looking ahead:

The next version of Lip-Wise promises even more impressive capabilities:

• Cutting-edge HD Wav2Lip models: I'm currently training state-of-the-art HD Wav2Lip models designed to further enhance the realism and precision of lip synchronization.
• Dedicated video super-resolution models: To unlock the full potential of video content, I'm actively integrating advanced video super-resolution models. These models will go beyond the limitations of image super-resolution by utilizing frame-to-frame context for exceptional video fidelity.

I'm dedicated to continually pushing the boundaries of lip-syncing technology with Lip-Wise. Stay tuned for future updates as I strive to empower your creativity with groundbreaking solutions.