DEDUPED Hot AI repos, turned into use cases Get 5 a day

What you can do with Diffusion OPSD

DiffusionOPSD is an external Python implementation of an on-policy self-distillation method for reward-guided post-training of diffusion models such as SD3.5-M and Z-Image-Turbo.

Trended earlier

worldbench/DiffusionOPSD on GitHub

Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk.

What do you want to do?

3 use cases

  • Post-train SD3.5-M toward a chosen image reward

    Fine-tune SD3.5-M with Diffusion OPSD using a reward preset and measure the held-out score.

    View prompt
    Using the repo worldbench/DiffusionOPSD, set up the environment and run the SD3.5-M training preset for <reward name>. Report the training command used, the reward checkpoint path, and the final evaluation score for <reward name>. Verify that scripts/download_reward_weights.sh completed without errors before starting training.

  • Verify reward models load before a long training run

    Run the README setup checks and gradient smoke tests so a failed reward stack does not waste GPU hours.

    View prompt
    In the repo worldbench/DiffusionOPSD, run the reward setup checks for <reward name> with the SD3.5 backbone and the gradient smoke test. Report which commands passed or failed and the error text for any failure. Verify that the smoke test ran the differentiable reward path before reporting success.

  • Train the Z-Image-Turbo backbone at 1024 resolution

    Run Diffusion OPSD on the few-step Z-Image-Turbo regime at 1024 pixels.

    View prompt
    In the repo worldbench/DiffusionOPSD, confirm ZImagePipeline imports in the installed Diffusers, then start the Z-Image-Turbo training preset for <reward name>. Report the import check result, the launch command, and the first logged seconds per update. Verify that the import check passed before launching training.

How to use Diffusion OPSD

  1. Check you have what it needs: A recent Linux machine with CUDA GPUs (eight-GPU profiles are reported in the README); Python 3.10 or 3.11 and conda; A Hugging Face account that has accepted the SD3.5-M license, for the gated model.
  2. Open worldbench/DiffusionOPSD on GitHub and read the README and license.
  3. Run git clone https://github.com/worldbench/DiffusionOPSD.git and cd DiffusionOPSD.
  4. Run conda create -n diffusionopsd python=3.11 -y and conda activate diffusionopsd.
  5. Run pip install -e ".[rewards]" and then pip install --no-deps 'image-reward==1.5'.
  6. Set export REWARD_CKPT_PATH="$PWD/reward_ckpts" and run bash scripts/download_reward_weights.sh.

These steps come from our top use case, Post-train SD3.5-M toward a chosen image reward, which also has a copy-paste agent prompt.

Diffusion OPSD alternatives · Best AI repos for Research

About this repo

Requirements

  • A recent Linux machine with CUDA GPUs (eight-GPU profiles are reported in the README)
  • Python 3.10 or 3.11 and conda
  • A Hugging Face account that has accepted the SD3.5-M license, for the gated model
  • Enough disk space for model weights, reward checkpoints, and the Pick-a-Pic prompt manifest
Maintain Diffusion OPSD? Add the Featured on Deduped badge

Paste this into your README:

[![Featured on Deduped](https://dedupedai.com/badge/worldbench-diffusionopsd.svg)](https://dedupedai.com/use-cases/worldbench-diffusionopsd/?ref=badge)

More badge styles and HTML

Questions about Diffusion OPSD

What is Diffusion OPSD?

DiffusionOPSD is an external Python implementation of an on-policy self-distillation method for reward-guided post-training of diffusion models such as SD3.5-M and Z-Image-Turbo. It is an open-source project on GitHub: https://github.com/worldbench/DiffusionOPSD.

Who is Diffusion OPSD for?

Built for ML researchers and engineers with CUDA GPUs who want to post-train diffusion image models against image reward scores.

What can I do with Diffusion OPSD?

3 use cases so far, for example: Post-train SD3.5-M toward a chosen image reward; Verify reward models load before a long training run; Train the Z-Image-Turbo backbone at 1024 resolution. Each has steps and a copy-paste agent prompt.

What do I need to run Diffusion OPSD?

A recent Linux machine with CUDA GPUs (eight-GPU profiles are reported in the README); Python 3.10 or 3.11 and conda; A Hugging Face account that has accepted the SD3.5-M license, for the gated model; Enough disk space for model weights, reward checkpoints, and the Pick-a-Pic prompt manifest.

What are alternatives to Diffusion OPSD?

Other hot AI repos in Research: Scientific Agent Skills, Speaker Mem R1, DSH Council, Evo Bot, AI Engineering Course.

Is Diffusion OPSD safe to run?

Deduped does not own, audit or vouch for it. Read the README and license, run it on your own systems first and review every command before you run it.