Pinned Loading

1. AlignProp uses direct reward backpropogation for the alignment of large-scale text-to-image diffusion models. Our method is 25x more sample and compute efficient than reinforcement learning methods…

   Python 326 11

2. Video Diffusion Alignment via Reward Gradients. We improve a variety of video diffusion models such as VideoCrafter, OpenSora, ModelScope and StableVideoDiffusion by finetuning them using various r…

   Python 318 15

3. Sim2Reason: Solving Physics Olympiad via Reinforcement Learning on Physics Simulators. We present a method for turning physics simulators into scalable generators of question–answer pairs for impro…

   Python 174 24

4. UniDisc: A discrete diffusion model for joint multimodal generation, enabling controllable and efficient text-image synthesis, editing, and inpainting.

   Python 142 6

5. Official PyTorch implementation and models for paper "Diffusion Beats Autoregressive in Data-Constrained Settings". We find diffusion models are significantly more data-efficient than standard left…

   Python 128 5

6. Diffusion-TTA improves pre-trained discriminative models such as image classifiers or segmentors using pre-trained generative models.

   Python 80 5