INSTRUCTOR-Large Open-source General Embedding Model - Supports Instruction Control for Text Fragment Vector Mapping

Instructor Large

Developed by hku-nlp

INSTRUCTOR-Large is a general embedding model capable of mapping any text fragment into a fixed-length vector, achieving domain specificity and task awareness through instruction control.

Text Embedding

Transformers

EnglishOpen Source License:Apache-2.0 #Instruction-driven Embedding #Domain Adaptation #Task-aware Vectors

Downloads 66

Release Time : 12/17/2022

Model Overview

This model can map any text fragment (such as titles, sentences, documents, etc.) into a fixed-length vector during the testing phase without requiring additional training. Through instruction control, the generated embedding vectors exhibit domain specificity (e.g., optimized for fields like science or finance) and task awareness (e.g., tailored for tasks like classification or information retrieval).

Model Features

Instruction-controlled Embedding

Generates embedding vectors through instruction control, making them domain-specific and task-aware.

General Text Embedding

Capable of processing any type of text fragment, including titles, sentences, and documents.

No Additional Training Required

Generates customized embedding vectors during the testing phase without requiring additional training.

Model Capabilities

Text Embedding

Sentence Similarity Calculation

Feature Extraction

Use Cases

Information Retrieval

Scientific Literature Retrieval

Improves the accuracy of scientific literature retrieval by generating domain-specific embedding vectors for science through instruction control.

Text Classification

Financial News Classification

Generates finance-specific embedding vectors through instruction control for financial news classification tasks.

🚀 hku-nlp/instructor-large

This is a general embedding model that maps any text to a fixed-length vector at test time without further training. With instructions, the embeddings are domain-specific and task-aware.

🚀 Quick Start

This model is a general embedding model. It can map any text (such as a title, a sentence, a document, etc.) to a fixed-length vector during testing without further training. With instructions, the embeddings are domain-specific (e.g., specialized for science, finance, etc.) and task-aware (e.g., customized for classification, information retrieval, etc.). The model is easy to use with the sentence-transformer library.

✨ Features

Maps any text to a fixed - length vector without further training.
Produces domain - specific and task - aware embeddings with instructions.

📦 Installation

git clone https://github.com/HKUNLP/instructor-embedding
cd sentence-transformers
pip install -e .

💻 Usage Examples

Basic Usage

You can use the model to calculate domain - specific and task - aware embeddings as follows:

from sentence_transformers import SentenceTransformer
sentence = "3D ActionSLAM: wearable person tracking in multi-floor environments"
instruction = "Represent the Science title; Input:"
model = SentenceTransformer('hku-nlp/instructor-large')
embeddings = model.encode([[instruction,sentence,0]])
print(embeddings)

Advanced Usage

You can further use the model to compute similarities between two groups of sentences, with customized embeddings:

from sklearn.metrics.pairwise import cosine_similarity
sentences_a = [['Represent the Science sentence; Input: ','Parton energy loss in QCD matter',0], 
               ['Represent the Financial statement; Input: ','The Federal Reserve on Wednesday raised its benchmark interest rate.',0]]
sentences_b = [['Represent the Science sentence; Input: ','The Chiral Phase Transition in Dissipative Dynamics', 0],
               ['Represent the Financial statement; Input: ','The funds rose less than 0.5 per cent on Friday',0]]
embeddings_a = model.encode(sentences_a)
embeddings_b = model.encode(sentences_b)
similarities = cosine_similarity(embeddings_a,embeddings_b)
print(similarities)

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご