958 views
Embeddings, or vectors, are a fundamental part of generative artificial intelligence applications. In this video, we will see what your objective is, how to generate them the hard way and the easy way and, in the end, how to use a database with Vector Search support to find similarities. Test DataStax Astra for Vector Search: https://bit.ly/41jSOS5 Ranking of models for embedding in Hugging Face: https://huggingface.co/spaces/mteb/le... Notebook: https://github. com/smatioli/astra-lan... Linkedin: / samuelmatioli 0:00 Home 0:15 What are embeddings 0:55 Vector Search 1:24 Tokens and Tokenization 2:45 Models for embedding in HuggingFace 4:05 Starting the generation of embedding 5:10 Return of the embedding model 5:28 Pooling 6:10 Finishing the embedding with normalization 6:35 Embeddings in easy mode with Sentence Transformers 7:34 Storing vectors in Cassandra 7:55 Library Python CassIO 8:13 Storing the vectors 9:05 Data model for vectors in Cassandra 9:40 Loading the vectors 10:25 Searching content for similarity 12:00 Conclusion #IAGenerativa #Cassandra #VectorSearch #Embedding #Python