langchain/docs/docs/integrations/text_embedding/sentence_transformers.ipynb

135 lines
3.2 KiB
Plaintext

{
"cells": [
{
"cell_type": "markdown",
"id": "ed47bb62",
"metadata": {},
"source": [
"# Sentence Transformers on Hugging Face\n",
"\n",
">[Hugging Face sentence-transformers](https://huggingface.co/sentence-transformers) is a Python framework for state-of-the-art sentence, text and image embeddings.\n",
">You can use these embedding models from the `HuggingFaceEmbeddings` class.\n",
"\n",
":::caution\n",
"\n",
"Running sentence-transformers locally can be affected by your operating system and other global factors. It is recommended for experienced users only.\n",
"\n",
":::\n",
"\n",
"## Setup\n",
"\n",
"You'll need to install the `langchain_huggingface` package as a dependency:"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "06c9f47d",
"metadata": {},
"outputs": [],
"source": [
"%pip install -qU langchain-huggingface"
]
},
{
"cell_type": "markdown",
"id": "8fb16f74",
"metadata": {},
"source": [
"## Usage"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "ff9be586",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[-0.038338568061590195, 0.12346471101045609, -0.028642969205975533, 0.05365273356437683, 0.008845377...\n"
]
}
],
"source": [
"from langchain_huggingface import HuggingFaceEmbeddings\n",
"\n",
"embeddings = HuggingFaceEmbeddings(model_name=\"all-MiniLM-L6-v2\")\n",
"\n",
"text = \"This is a test document.\"\n",
"query_result = embeddings.embed_query(text)\n",
"\n",
"# show only the first 100 characters of the stringified vector\n",
"print(str(query_result)[:100] + \"...\")"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "bb5e74c0",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[[-0.038338497281074524, 0.12346471846103668, -0.028642890974879265, 0.05365274101495743, 0.00884535...\n"
]
}
],
"source": [
"doc_result = embeddings.embed_documents([text, \"This is not a test document.\"])\n",
"print(str(doc_result)[:100] + \"...\")"
]
},
{
"cell_type": "markdown",
"id": "1e6525cb",
"metadata": {},
"source": [
"## Troubleshooting\n",
"\n",
"If you are having issues with the `accelerate` package not being found or failing to import, installing/upgrading it may help:"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bbae70f7",
"metadata": {},
"outputs": [],
"source": [
"%pip install -qU accelerate"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.5"
},
"vscode": {
"interpreter": {
"hash": "7377c2ccc78bc62c2683122d48c8cd1fb85a53850a1b1fc29736ed39852c9885"
}
}
},
"nbformat": 4,
"nbformat_minor": 5
}