diff --git a/site/en/embeddings/embed-with-text-embeddings-inference.md b/site/en/embeddings/embed-with-text-embeddings-inference.md new file mode 100644 index 000000000..207d39bac --- /dev/null +++ b/site/en/embeddings/embed-with-text-embeddings-inference.md @@ -0,0 +1,102 @@ +--- +id: embed-with-text-embeddings-inference.md +order: 14 +summary: This article demonstrates how to use text-embeddings-inference in Milvus to encode documents and queries into dense vectors. +title: text-embeddings-inference(TEI) +--- + +# text-embeddings-inference(TEI) + +Text Embeddings Inference (TEI) is a comprehensive toolkit designed for efficient deployment and serving of open source text embeddings models. It enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE, and E5. + +Key Features: +- Streamlined Deployment +- Efficient Resource Utilization +- Dynamic Batching +- Optimized Inference +- Safetensors weight loading +- Production-Ready + +💡 **Tip:** Ensure that the TEI service is running before proceeding. TEI Deployment Reference Documentation https://github.com/huggingface/text-embeddings-inference + +To use this feature, install the necessary dependencies: + +```bash +pip install --upgrade pymilvus +pip install "pymilvus[model]" +``` + +Then, instantiate the __TEIEmbeddingFunction__: + +```python +from pymilvus import model + +tei_ef = model.dense.TEIEmbeddingFunction( + api_url='http://127.0.0.1:8000' +) +``` + +__Parameters__: + +- __api_url__ (_string_) + + TEI deployment api address. + + +To create embeddings for documents, use the __encode_documents()__ method: + +```python +docs = [ + "Artificial intelligence was founded as an academic discipline in 1956.", + "Alan Turing was the first person to conduct substantial research in AI.", + "Born in Maida Vale, London, Turing was raised in southern England.", +] + +docs_embeddings = tei_ef.encode_documents(docs) + +# Print embeddings +print("Embeddings:", docs_embeddings) +# Print dimension and shape of embeddings +print("Dim:", tei_ef.dim, docs_embeddings[0].shape) +``` + +The expected output is similar to the following: + +```python +Embeddings: [array([-3.09392996e-02, -1.80662833e-02, 1.34775648e-02, 2.77156215e-02, + -4.86349640e-03, -3.12581174e-02, -3.55921760e-02, 5.76934684e-03, + 2.80773244e-03, 1.35783911e-01, 3.59678417e-02, 6.17732145e-02, +... + -4.61330153e-02, -4.85207550e-02, 3.13997865e-02, 7.82178566e-02, + -4.75336798e-02, 5.21207601e-02, 9.04406682e-02, -5.36676683e-02], + dtype=float32)] +Dim: 512 (512,) +``` + +To create embeddings for queries, use the __encode_queries()__ method: + +```python +queries = ["When was artificial intelligence founded", + "Where was Alan Turing born?"] + +query_embeddings = tei_ef.encode_queries(queries) + +# Print embeddings +print("Embeddings:", query_embeddings) +# Print dimension and shape of embeddings +print("Dim:", tei_ef.dim, query_embeddings[0].shape) +``` + +The expected output is similar to the following: + +```python +Embeddings: [array([-2.52114702e-02, -5.29330298e-02, 1.14570223e-02, 1.95571519e-02, + -2.46500354e-02, -2.66519729e-02, -8.48201662e-03, 2.82961670e-02, + -3.65092754e-02, 7.50745758e-02, 4.28900979e-02, 7.18822703e-02, +... + -6.76431581e-02, -6.45996556e-02, -4.67132553e-02, 4.78532910e-02, + -2.31596199e-03, 4.13446948e-02, 1.06935494e-01, -1.08258888e-01], + dtype=float32)] +Dim: 512 (512,) +``` + diff --git a/site/en/rerankers/rerankers-text-embeddings-inference.md b/site/en/rerankers/rerankers-text-embeddings-inference.md new file mode 100644 index 000000000..7b7bfba47 --- /dev/null +++ b/site/en/rerankers/rerankers-text-embeddings-inference.md @@ -0,0 +1,83 @@ +--- +id: rerankers-text-embeddings-inference.md +order: 7 +summary: Milvus supports the deployment of open source reranker models by TEI through the “TEIRerankFunction” class. This function allows you to efficiently score the relevance of query-document pairs. +title: text-embeddings-inference(TEI) - Rerankers +--- + +# text-embeddings-inference(TEI) + +Text Embeddings Inference (TEI) is a comprehensive toolkit designed for efficient deployment and serving of open source text embeddings models. It enables high-performance extraction for the most popular models, including bge-reranker-large, roberta-base-go_emotions. + +Key Features: +- Streamlined Deployment +- Efficient Resource Utilization +- Dynamic Batching +- Optimized Inference +- Safetensors weight loading +- Production-Ready + +💡 **Tip:** Ensure that the TEI service is running before proceeding. TEI Deployment Reference Documentation https://github.com/huggingface/text-embeddings-inference + +To use this feature, install the necessary dependencies: + +```bash +pip install --upgrade pymilvus +pip install "pymilvus[model]" +``` + +Then, instantiate the `TEIRerankFunction`: + +```python +from pymilvus.model.reranker import TEIRerankFunction + +tei_rf = TEIRerankFunction( + api_url='http://127.0.0.1:8000' +) +``` + +__Parameters__: + +- __api_url__ (_string_) + + TEI deployment api address. + +Then, use the following code to rerank documents based on the query: + +```python +query = "What event in 1956 marked the official birth of artificial intelligence as a discipline?" + +documents = [ + "In 1950, Alan Turing published his seminal paper, 'Computing Machinery and Intelligence,' proposing the Turing Test as a criterion of intelligence, a foundational concept in the philosophy and development of artificial intelligence.", + "The Dartmouth Conference in 1956 is considered the birthplace of artificial intelligence as a field; here, John McCarthy and others coined the term 'artificial intelligence' and laid out its basic goals.", + "In 1951, British mathematician and computer scientist Alan Turing also developed the first program designed to play chess, demonstrating an early example of AI in game strategy.", + "The invention of the Logic Theorist by Allen Newell, Herbert A. Simon, and Cliff Shaw in 1955 marked the creation of the first true AI program, which was capable of solving logic problems, akin to proving mathematical theorems." +] + +results = tei_rf( + query=query, + documents=documents, + top_k=3, +) + +for result in results: + print(f"Index: {result.index}") + print(f"Score: {result.score:.6f}") + print(f"Text: {result.text}\n") +``` + +The expected output is similar to the following: + +```python +Index: 1 +Score: 0.9971661 +Text: The Dartmouth Conference in 1956 is considered the birthplace of artificial intelligence as a field; here, John McCarthy and others coined the term 'artificial intelligence' and laid out its basic goals. + +Index: 2 +Score: 0.00809329 +Text: In 1951, British mathematician and computer scientist Alan Turing also developed the first program designed to play chess, demonstrating an early example of AI in game strategy. + +Index: 0 +Score: 0.002491968 +Text: The invention of the Logic Theorist by Allen Newell, Herbert A. Simon, and Cliff Shaw in 1955 marked the creation of the first true AI program, which was capable of solving logic problems, akin to proving mathematical theorems. +```