{"payload":{"pageCount":1,"repositories":[{"type":"Public","name":"SUR-adapter","owner":"Qrange-group","isFork":false,"description":"ACM MM'23 (oral), SUR-adapter for pre-trained diffusion models can acquire the powerful semantic understanding and reasoning capabilities from large language models to build a high-quality textual semantic representation for text-to-image generation.","allTopics":["adapter","pytorch","image-generation","knowledge-distillation","diffusion-models","large-language-models"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":7,"starsCount":105,"forksCount":2,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2024-04-24T03:09:31.163Z"}},{"type":"Public","name":"Mirror-Gradient","owner":"Qrange-group","isFork":false,"description":"WWW'24, Mirror Gradient (MG) makes multimodal recommendation models approach flat local minima easier compared to models with normal training.","allTopics":["recommender-system","recommendation","multimodal","gradient-method","flat-local-minima"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":0,"starsCount":9,"forksCount":1,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2024-02-22T03:16:26.864Z"}},{"type":"Public","name":"SEM","owner":"Qrange-group","isFork":false,"description":"SEM can automatically decide to select and integrate attention operators to compute attention maps.","allTopics":["pytorch","image-classification","self-attention","excitation-networks"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":0,"starsCount":8,"forksCount":2,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2023-06-16T02:52:09.976Z"}},{"type":"Public","name":"SPEM","owner":"Qrange-group","isFork":false,"description":"MMM'23, SPEM adopts a self-adaptive pooling strategy based on global max-pooling, global min-pooling and a lightweight module for producing the attention map.","allTopics":["pytorch","image-classification","pooling","self-attention"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":0,"starsCount":2,"forksCount":0,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2023-06-16T02:51:36.485Z"}},{"type":"Public","name":"LSAS","owner":"Qrange-group","isFork":false,"description":"ICME'23, Lightweight sub-attention strategy (LSAS) utilizes high-order sub-attention modules to improve the original self-attention modules. ","allTopics":["pytorch","image-classification","self-attention","sub-attention","attention-bias"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":0,"starsCount":3,"forksCount":0,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2023-06-16T02:51:18.270Z"}},{"type":"Public","name":"CEM","owner":"Qrange-group","isFork":false,"description":"EMNLP'22, CEM improves MHCH performance by correcting prediction bias and training an auxiliary cost simulator based on user state and labor cost causal graph, without requiring complex model crafting.","allTopics":["python","natural-language-processing","tensorflow","causal-inference","dialogue-systems","counterfactual","causal-graph"],"primaryLanguage":{"name":"Python","color":"#3572A5"},"pullRequestCount":0,"issueCount":0,"starsCount":11,"forksCount":0,"license":"MIT License","participation":null,"lastUpdated":{"hasBeenPushedTo":true,"timestamp":"2022-10-09T03:13:30.304Z"}}],"repositoryCount":6,"userInfo":null,"searchable":true,"definitions":[],"typeFilters":[{"id":"all","text":"All"},{"id":"public","text":"Public"},{"id":"source","text":"Sources"},{"id":"fork","text":"Forks"},{"id":"archived","text":"Archived"},{"id":"template","text":"Templates"}],"compactMode":false},"title":"Repositories"}