1  ---
  2  title: "FastembedTextEmbedder"
  3  id: fastembedtextembedder
  4  slug: "/fastembedtextembedder"
  5  description: "This component computes the embeddings of a string using embedding models supported by FastEmbed."
  6  ---
  7  
  8  # FastembedTextEmbedder
  9  
 10  This component computes the embeddings of a string using embedding models supported by FastEmbed.
 11  
 12  |                                        |                                                                                             |
 13  | :------------------------------------- | :------------------------------------------------------------------------------------------ |
 14  | **Most common position in a pipeline** | Before an embedding [Retriever](../retrievers.mdx)  in a query/RAG pipeline                  |
 15  | **Mandatory run variables**            | “text”: A string                                                                            |
 16  | **Output variables**                   | “embedding”: A vector (list of float numbers)                                               |
 17  | **API reference**                      | [FastEmbed](/reference/fastembed-embedders)                                                        |
 18  | **GitHub link**                        | https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/fastembed |
 19  
 20  This component should be used to embed a simple string (such as a query) into a vector. For embedding lists of documents, use the [`FastembedDocumentEmbedder`](fastembeddocumentembedder.mdx), which enriches the document with the computed embedding, known as vector.
 21  
 22  ## Overview
 23  
 24  `FastembedTextEmbedder` transforms a string into a vector that captures its semantics using embedding [models supported by FastEmbed](https://qdrant.github.io/fastembed/examples/Supported_Models/).
 25  
 26  When you perform embedding retrieval, use this component first to transform your query into a vector. Then, the embedding Retriever will use the vector to search for similar or relevant documents.
 27  
 28  ### Compatible models
 29  
 30  You can find the original models in the [FastEmbed documentation](https://qdrant.github.io/fastembed/).
 31  
 32  Currently, most of the models in the [Massive Text Embedding Benchmark (MTEB) Leaderboard](https://huggingface.co/spaces/mteb/leaderboard) are compatible with FastEmbed. You can look for compatibility in the [supported model list](https://qdrant.github.io/fastembed/examples/Supported_Models/).
 33  
 34  ### Installation
 35  
 36  To start using this integration with Haystack, install the package with:
 37  
 38  ```bash
 39  pip install fastembed-haystack
 40  ```
 41  
 42  ### Instructions
 43  
 44  Some recent models that you can find in MTEB require prepending the text with an instruction to work better for retrieval.
 45  For example, if you use `[BAAI/bge-large-en-v1.5](https://huggingface.co/BAAI/bge-large-en-v1.5#model-list)` model, you should prefix your query with the `instruction: “passage:”`.
 46  
 47  This is how it works with `FastembedTextEmbedder`:
 48  
 49  ```python
 50  instruction = "passage:"
 51  embedder = FastembedTextEmbedder(
 52  	*model="*BAAI/bge-large-en-v1.5",
 53  	prefix=instruction)
 54  ```
 55  
 56  ### Parameters
 57  
 58  You can set the path where the model will be stored in a cache directory. Also, you can set the number of threads a single `onnxruntime` session can use.
 59  
 60  ```python
 61  cache_dir= "/your_cacheDirectory"
 62  embedder = FastembedTextEmbedder(
 63  	*model="*BAAI/bge-large-en-v1.5",
 64  	cache_dir=cache_dir,
 65  	threads=2
 66  )
 67  ```
 68  
 69  If you want to use the data parallel encoding, you can set the parameters `parallel` and `batch_size`.
 70  
 71  - If parallel > 1, data-parallel encoding will be used. This is recommended for offline encoding of large datasets.
 72  - If parallel is 0, use all available cores.
 73  - If None, don't use data-parallel processing; use default `onnxruntime` threading instead.
 74  
 75  :::tip
 76  If you create a Text Embedder and a Document Embedder based on the same model, Haystack uses the same resource behind the scenes to save resources.
 77  
 78  :::
 79  
 80  ## Usage
 81  
 82  ### On its own
 83  
 84  ```python
 85  from haystack_integrations.components.embedders.fastembed import FastembedTextEmbedder
 86  
 87  text = """It clearly says online this will work on a Mac OS system.
 88  The disk comes and it does not, only Windows.
 89  Do Not order this if you have a Mac!!"""
 90  text_embedder = FastembedTextEmbedder(model="BAAI/bge-small-en-v1.5")
 91  text_embedder.warm_up()
 92  embedding = text_embedder.run(text)["embedding"]
 93  ```
 94  
 95  ### In a pipeline
 96  
 97  ```python
 98  from haystack import Document, Pipeline
 99  from haystack.components.retrievers.in_memory import InMemoryEmbeddingRetriever
100  from haystack.document_stores.in_memory import InMemoryDocumentStore
101  from haystack_integrations.components.embedders.fastembed import (
102      FastembedDocumentEmbedder,
103      FastembedTextEmbedder,
104  )
105  
106  document_store = InMemoryDocumentStore(embedding_similarity_function="cosine")
107  
108  documents = [
109      Document(content="My name is Wolfgang and I live in Berlin"),
110      Document(content="I saw a black horse running"),
111      Document(content="Germany has many big cities"),
112      Document(content="fastembed is supported by and maintained by Qdrant."),
113  ]
114  
115  document_embedder = FastembedDocumentEmbedder()
116  document_embedder.warm_up()
117  documents_with_embeddings = document_embedder.run(documents)["documents"]
118  document_store.write_documents(documents_with_embeddings)
119  
120  query_pipeline = Pipeline()
121  query_pipeline.add_component("text_embedder", FastembedTextEmbedder())
122  query_pipeline.add_component(
123      "retriever",
124      InMemoryEmbeddingRetriever(document_store=document_store),
125  )
126  query_pipeline.connect("text_embedder.embedding", "retriever.query_embedding")
127  
128  query = "Who supports FastEmbed?"
129  
130  result = query_pipeline.run({"text_embedder": {"text": query}})
131  
132  print(result["retriever"]["documents"][0])  # noqa: T201
133  
134  ## Document(id=...,
135  ##  content: 'FastEmbed is supported by and maintained by Qdrant.',
136  ##  score: 0.758..)
137  ```
138  
139  ## Additional References
140  
141  🧑‍🍳 Cookbook: [RAG Pipeline Using FastEmbed for Embeddings Generation](https://haystack.deepset.ai/cookbook/rag_fastembed)