Learn about amdgpu OpenCL neural network use for video enhancing
an invention by tjyrinki_suse
Updated over 3 years ago. 2 hacker ♥️. 1 follower.

Project Description

I've had a fairly modern AMD GPU for some time, but I haven't had time to check the current status of neural network based video editing. I know many are proprietary CUDA specific, sadly, even if projects themselves would be open source, but I'd like to check which ones would support OpenCL as well or any other means to use an AMD GPU with fully open source stack (including drivers).

Also update current status of the mess with OpenCL 1.2 / 2.0 / 3.0, ie is 3.0 becoming useful to eventually replace 1.2.

Goal for this Hackweek

Have an overview, possibly some example videos to post if something functional found.

Resources

Any and all tips specific to video enhancing (superresolution, motion interpolation etc) hardware accelerated with AMD GPU on openSUSE welcome, as is discussion.

Looking for hackers with the skills:

opencl amd gpu vulkan videoprocessing

This project is part of:

Hack Week 20

Activity

  • over 3 years ago: tjyrinki_suse added keyword "opencl" to this project.
  • over 3 years ago: tjyrinki_suse added keyword "amd" to this project.
  • over 3 years ago: tjyrinki_suse added keyword "gpu" to this project.
  • over 3 years ago: tjyrinki_suse added keyword "vulkan" to this project.
  • over 3 years ago: tjyrinki_suse added keyword "videoprocessing" to this project.
  • over 3 years ago: okurz liked this project.
  • over 3 years ago: tjyrinki_suse started this project.
  • over 3 years ago: tjyrinki_suse liked this project.
  • over 3 years ago: tjyrinki_suse originated this project.


    • Comments

    • tjyrinki_suse
      over 3 years ago by tjyrinki_suse | Reply

      Blogged about my learnings at Super Resolution Video Enhancing with AMD GPU

    Similar Projects

    COOTWbot by ngetahun

    Project Description

    At SCC, we have a rotating task of COOTW (Commanding Office of the Week). This task involves responding to customer requests from jira and slack help channels, monitoring production systems and doing small chores. Usually, we have documentation to help the COOTW answer questions and quickly find fixes. Most of these are distributed across github, trello and SUSE Support documentation. The aim of this project is to explore the magic of LLMs and create a conversational bot.

    Goal for this Hackweek

    • Evaluate data gathering and cleanup. Create sensible embeddings to be used by LLMs.
    • Explore performance of Open-source llama based models (LLama, Vicuna, Mistral) in generating coherent and reasonably fast responses.
    • Look into creating evaluation data for later use
    • [Optional][requires beefy GPUs] Train Low Rank Adaptations (LoRA) and compare results.

    Resources