Views
No views yet
intextus-embed runtime:pip install intextus-embed1from intextus import IntextusEncoder, compute_maxsim
2
3# Automatically downloads and caches the model from Hugging Face
4model = IntextusEncoder("mxbai-edge-colbert-v0-17m")
5
6# Encode queries and documents
7query_embeddings = model.encode_queries("What is ultra-low latency?")
8doc_embeddings = model.encode_docs("ONNX runtime bypasses the PyTorch layer completely.")
9
10# Compute the MaxSim similarity score via NumPy
11score = compute_maxsim(query_embeddings[0], doc_embeddings[0])
12print(f"Relevance Score (MaxSim): {score:.4f}")