[1/3] Want to capture a fantastic HDR image by 2 simple shots with your cellphone? Try UltraFusion HDR. It takes two images with exposure differences up to 9 stops, and it robustly generates HDR output. Try your own captured images (supporting 4Kx3K): huggingface.co/spaces/iimmort…

Again guys, this is not our Wenfeng.

One interesting learning from the R1 and K1.5 tech report is the usage of string matching based binary reward: I’ve tried it myself in 2022 using FlanT5, my friends tried it in 2023 with Llama 1 and in early 2024 with llama 2, but all failed completely. It is only after late

🥳Update for <Improving Video Generation with Human Feedback>: - The Reward Model for video has been released: github.com/KwaiVGI/VideoA… huggingface.co/KwaiVGI/VideoR… - together with VideoGen-RewardBench dataset: huggingface.co/datasets/KwaiV…

Nice work👏👏👏

Thanks for sharing your work! If your generative model uses flow-matching, feel free to try our Flow-GRPO. Our code is at: github.com/yifan123/flow_…

Thanks for sharing your work! Flow-GRPO significantly enhances the capabilities of flow matching models within a given dimensionality. Our code is at: github.com/yifan123/flow_…

Really happy to see that Long-RL adopted Flow-GRPO (github.com/yifan123/flow_…) and implemented it on the popular RL repo verl. Amazing work 👏

Super excited to see Qwen-Image using Flow-GRPO for post-training! As a purely method-focused academic repo, it has already passed 1k stars 🎉. Several recent community improvements have also significantly boosted its performance and training speed!