diffusers Diffusers 0.43.0: MiniMax-H3-Image, Pika H3 Flex, P-H3-Image, Kandinsky S2V, and PAN 2

Version 0.43.0 of diffusers introduces five new modular pipelines for advanced image, video, and simulation tasks, including text-to-image generation with reference images, speech-driven animation, and scene simulation. Key improvements include enhanced attention backends, better LoRA support, and optimized checkpoint saving for production use.