Search results

Question about training timestep setting in multi-reward training

Status: Open.

#44 In NVlabs/DiffusionNFT;

· MaoMao78777 opened on Feb 12, 2026

About training efficiency

Status: Open.

#43 In NVlabs/DiffusionNFT;

· shinetzh opened on Feb 12, 2026

timesteps bug

Status: Open.

#42 In NVlabs/DiffusionNFT;

· byliutao opened on Feb 11, 2026

Does this framework currently only support SD3.5?

Status: Open.

#39 In NVlabs/DiffusionNFT;

· BoweiLiu628 opened on Feb 8, 2026

Questions about training

Status: Open.

#37 In NVlabs/DiffusionNFT;

· LAW1223 opened on Jan 16, 2026

Question about the specific pipelines of Multi-Reward joint training.

Status: Open.

#36 In NVlabs/DiffusionNFT;

· Fuzm7 opened on Jan 13, 2026

On the issue that reproducing your open-source code yields incorrect training results

Status: Open.

#22 In NVlabs/DiffusionNFT;

· Nnnny-a opened on Oct 24, 2025

I am trying to reproduce your result on SDXL(ddpm method not flow matching)

Status: Open.

#12 In NVlabs/DiffusionNFT;

· shoutOutYangJie opened on Oct 15, 2025

xFormers can't load C++/CUDA extensions

Status: Open.

#7 In NVlabs/DiffusionNFT;

· jaylee2000 opened on Oct 2, 2025

Qualitative results trained on single reward / Head-to-head comparison with multiple rewards

Status: Open.

#6 In NVlabs/DiffusionNFT;

· jaylee2000 opened on Oct 1, 2025