The post says dropping a modality makes audio and video learn a common representation without class labels.
A post introduces LeAVJEPA as a self-supervised approach to learning from sound and vision. It describes one shared encoder trained without class labels and says dropping a modality makes audio and video learn a common representation.
20.9K
2 Sources, first seen ago
—
Not ranked yet
—
Not ranked yet
@arnosolin🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class labels. Dropping a modality makes audio & video learn a common representation. 🧵
@ylecunRT @arnosolin: 🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class…@arnosolin🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class labels. Dropping a modality makes audio & video learn a common representation. 🧵@ylecunRT @arnosolin: 🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class…@arnosolin🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class labels. Dropping a modality makes audio & video learn a common representation. 🧵@ylecunRT @arnosolin: 🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class…@arnosolin🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class labels. Dropping a modality makes audio & video learn a common representation. 🧵@ylecunRT @arnosolin: 🔥 #LeAVJEPA: A simple recipe for self-supervised learning from sound and vision 🔥 One shared encoder, trained without class…
