NomicEmbeddings
This will help you get started with Nomic embedding models using LangChain. For detailed documentation on NomicEmbeddings
features and configuration options, please refer to the API reference.
Overviewโ
Integration detailsโ
Provider | Package |
---|---|
Nomic | langchain-nomic |
Setupโ
To access Nomic embedding models you'll need to create a/an Nomic account, get an API key, and install the langchain-nomic
integration package.
Credentialsโ
Head to https://atlas.nomic.ai/ to sign up to Nomic and generate an API key. Once you've done this set the NOMIC_API_KEY
environment variable:
import getpass
import os
if not os.getenv("NOMIC_API_KEY"):
os.environ["NOMIC_API_KEY"] = getpass.getpass("Enter your Nomic API key: ")
If you want to get automated tracing of your model calls you can also set your LangSmith API key by uncommenting below:
# os.environ["LANGCHAIN_TRACING_V2"] = "true"
# os.environ["LANGCHAIN_API_KEY"] = getpass.getpass("Enter your LangSmith API key: ")
Installationโ
The LangChain Nomic integration lives in the langchain-nomic
package:
%pip install -qU langchain-nomic
Note: you may need to restart the kernel to use updated packages.
Instantiationโ
Now we can instantiate our model object and generate chat completions:
from langchain_nomic import NomicEmbeddings
embeddings = NomicEmbeddings(
model="nomic-embed-text-v1.5",
# dimensionality=256,
# Nomic's `nomic-embed-text-v1.5` model was [trained with Matryoshka learning](https://blog.nomic.ai/posts/nomic-embed-matryoshka)
# to enable variable-length embeddings with a single model.
# This means that you can specify the dimensionality of the embeddings at inference time.
# The model supports dimensionality from 64 to 768.
# inference_mode="remote",
# One of `remote`, `local` (Embed4All), or `dynamic` (automatic). Defaults to `remote`.
# api_key=... , # if using remote inference,
# device="cpu",
# The device to use for local embeddings. Choices include
# `cpu`, `gpu`, `nvidia`, `amd`, or a specific device name. See
# the docstring for `GPT4All.__init__` for more info. Typically
# defaults to CPU. Do not use on macOS.
)
Indexing and Retrievalโ
Embedding models are often used in retrieval-augmented generation (RAG) flows, both as part of indexing data as well as later retrieving it. For more detailed instructions, please see our RAG tutorials under the working with external knowledge tutorials.
Below, see how to index and retrieve data using the embeddings
object we initialized above. In this example, we will index and retrieve a sample document in the InMemoryVectorStore
.
# Create a vector store with a sample text
from langchain_core.vectorstores import InMemoryVectorStore
text = "LangChain is the framework for building context-aware reasoning applications"
vectorstore = InMemoryVectorStore.from_texts(
[text],
embedding=embeddings,
)
# Use the vectorstore as a retriever
retriever = vectorstore.as_retriever()
# Retrieve the most similar text
retrieved_documents = retriever.invoke("What is LangChain?")
# show the retrieved document's content
retrieved_documents[0].page_content
'LangChain is the framework for building context-aware reasoning applications'
Direct Usageโ
Under the hood, the vectorstore and retriever implementations are calling embeddings.embed_documents(...)
and embeddings.embed_query(...)
to create embeddings for the text(s) used in from_texts
and retrieval invoke
operations, respectively.
You can directly call these methods to get embeddings for your own use cases.
Embed single textsโ
You can embed single texts or documents with embed_query
:
single_vector = embeddings.embed_query(text)
print(str(single_vector)[:100]) # Show the first 100 characters of the vector
[0.024642944, 0.029083252, -0.14013672, -0.09082031, 0.058898926, -0.07489014, -0.0138168335, 0.0037