Wanyin App is an AI-based music sharing community with an intention to encourage music sharing and make music composition easier for music enthusiasts.
Wanyin’s library contains a massive amount of music uploaded by users. The primary task is to sort out the music of interest based on users’ previous behavior. We evaluated two classic models: user-based collaborative filtering (User-based CF) and item-based collaborative filtering (Item-based CF), as the potential recommender system models.
As its name suggests, searching for videos by image is the process of retrieving from the repository videos containing similar frames to the input image. One of the key steps is to turn videos into embeddings, which is to say, extract the key frames and convert their features to vectors. Now, some curious readers might wonder what the difference is between searching for video by image and searching for an image by image? In fact, searching for the key frames in videos is equivalent to searching for an image by image.
You can refer to our previous article Milvus x VGG: Building a Content-based Image Retrieval System if interested. …
This article deals with how Milvus implements the delete function. As a much-anticipated feature for many users, the delete function was introduced to Milvus v0.7.0. We did not call remove_ids in FAISS directly, instead, we came up with a brand new design to make deletion more efficient and support more index types.
In How Milvus Realizes Dynamic Data Update and Query, we introduced the entire process from inserting data to flushing data. Let’s recap on some of the basics. MemManager manages all insert buffers, with each MemTable corresponding to a collection (we renamed “table” to “collection” in Milvus v0.7.0). …