Search AI Products and News

A research team from New York University, MIT, and Google has recently proposed an innovative framework aimed at addressing the bottleneck of inference time scaling in diffusion models. This groundbreaking research transcends the traditional method of simply increasing the denoising steps, opening up new avenues for improving generative model performance. The framework primarily unfolds from two dimensions: one is leveraging feedback from validators, and the second is implementing algorithms to discover better noise candidates. The research team builds on a pre-trained SiT-XL model with a resolution of 256×256, maintaining 250 fixed denoising steps.

["Over 24 hours since the launch of Claude 3, it continues to reshape people's perceptions.", "Netizens believe that Claude 3 behaves as if it has a 'coherent subject' existence, unlike other large models.", "Some OpenAI users suggest that if Claude continues to showcase its capabilities, it might push OpenAI to release its next-generation model.", "Jimmy Apples releases the latest news, indicating that OpenAI's risk/reward assessment regarding the release of the next-generation model may be influenced by Claude 3."]

Good research does not fade away; it only becomes more relevant over time. In the past week, OpenAI's video generation model Sora has continued to attract attention. The team has launched a new version, SiT, which further improves quality, speed, and flexibility. The latest paper, "Scalable Diffusion Models with Transformers", has been recognized at ICCV2023, leading the advancement of technology. SiT explores a broader design space through interpolation, bringing new possibilities to the field of image generation.