Skip to main content

​
Overview

An instruction-finetuned text embedding model that can generate text embeddings tailored to any task (e.g., classification, retrieval, clustering, text evaluation, etc.) or domain (e.g., science, finance, etc.) by simply providing the task instruction in natural language.Takes customized text units (e.g. paragraph, sentence, document). Better performance than instructor-large.Learn about how to best use Instructor for specific tasks here.

​
Using the model

​
Installation:

!pip install transformers==4.20.0 InstructorEmbedding pinecone sentence-transformers

​
Create Index

from pinecone import Pinecone, ServerlessSpec

pc = Pinecone(api_key="API_KEY")

# Create Index
index_name = "instructor-xl"

if not pc.has_index(index_name):
    pc.create_index(
        name=index_name,
        dimension=1024,
        metric="cosine",
        spec=ServerlessSpec(
            cloud='aws',
            region='us-east-1'
        )
    )

index = pc.Index(index_name)

​
Embed & Upsert

# Embed data
data = [
    {"id": "vec1", "text": "Apple is a popular fruit known for its sweetness and crisp texture."},
    {"id": "vec2", "text": "The tech company Apple is known for its innovative products like the iPhone."},
    {"id": "vec3", "text": "Many people enjoy eating apples as a healthy snack."},
    {"id": "vec4", "text": "Apple Inc. has revolutionized the tech industry with its sleek designs and user-friendly interfaces."},
    {"id": "vec5", "text": "An apple a day keeps the doctor away, as the saying goes."},
]


# using Instructor, we need an instruction to append to passages

instruction = "Represent the following document for retrieval: "

from InstructorEmbedding import INSTRUCTOR

model = INSTRUCTOR('hkunlp/instructor-xl')


# align instructions with text data
# you can vary the instructions by data as well
instruction_embedding_pairs = [[instruction,  d["text"]] for d in data]

embeddings = model.encode(instruction_embedding_pairs)

vectors = []
for d, e in zip(data, embeddings):
    vectors.append({
        "id": d['id'],
        "values": e,
        "metadata": {'text': d['text']}
    })

index.upsert(
    vectors=vectors,
    namespace="ns1"
)

​
Query


query_instruction = "Represent this query for retrieving supporting documents: "

query = "Tell me about the tech company known as Apple"

x = model.encode([[query_instruction, query]])

results = index.query(
    namespace="ns1",
    vector=x[0].tolist(),
    top_k=3,
    include_values=False,
    include_metadata=True
)

print(results)
Lorem Ipsum