INSTRUCTOR Text Embedding Model Open-Source - Free for Sentence Similarity Calculation and Information Retrieval

Instructor Large Safetensors

Developed by gentlebowl

INSTRUCTOR is a text embedding model based on the T5 architecture, focusing on sentence similarity calculation and information retrieval tasks. It excels in various NLP tasks, including text classification, clustering, and semantic similarity evaluation.

Text Embedding

Transformers

EnglishOpen Source License:Apache-2.0 #Text Semantic Similarity #Multi-task Text Embedding #Information Retrieval Optimization

Downloads 16

Release Time : 4/25/2023

Model Overview

INSTRUCTOR is a powerful text embedding model capable of converting text into high-quality vector representations, suitable for various natural language processing tasks such as information retrieval, text classification, clustering, and semantic similarity calculation.

Model Features

Multi-task Support

Supports various NLP tasks, including sentence similarity, information retrieval, text classification, and clustering.

High Performance

Outperforms on multiple benchmark datasets such as MTEB and BEIR.

Flexible Embedding Vectors

Capable of generating high-quality text embedding vectors suitable for various downstream tasks.

Model Capabilities

Sentence similarity calculation

Information retrieval

Text classification

Text clustering

Semantic similarity evaluation

Prompt retrieval

Text re-ranking

Use Cases

Information Retrieval

Document Retrieval

Efficient document retrieval using INSTRUCTOR embedding vectors.

Achieves an average precision@10 of 38.1365 on the CQADupstack dataset.

Text Classification

Sentiment Analysis

Text sentiment classification using INSTRUCTOR.

Achieves an accuracy of 91.526% on the AmazonPolarity dataset.

Semantic Similarity

Sentence Similarity Calculation

Calculating semantic similarity between two sentences.

Achieves a Spearman correlation coefficient of 84.387 on the BIOSSES dataset.

🚀 INSTRUCTOR

INSTRUCTOR is a model designed for various text - related tasks such as sentence similarity, text classification, and information retrieval. It provides reliable performance across multiple datasets and tasks.

✨ Features

Diverse Task Support: Capable of handling tasks like Classification, Retrieval, Clustering, Reranking, and STS.
Multiple Datasets: Tested on a wide range of datasets including MTEB AmazonCounterfactualClassification, MTEB AmazonPolarityClassification, and many others.
Rich Metrics: Evaluated using metrics such as accuracy, ap, f1, map, mrr, etc.

📄 License

The model is licensed under the apache - 2.0 license.

📚 Documentation

Model Information

Property	Details
Model Name	INSTRUCTOR
Pipeline Tag	sentence - similarity
Tags	text - embedding, embeddings, information - retrieval, beir, text - classification, language - model, text - clustering, text - semantic - similarity, text - evaluation, prompt - retrieval, text - reranking, sentence - transformers, feature - extraction, sentence - similarity, transformers, t5, English, Sentence Similarity, natural_questions, ms_marco, fever, hotpot_qa, mteb
Language	en
Inference	false

Model Results

Classification Tasks

Dataset	Accuracy	AP	F1
MTEB AmazonCounterfactualClassification (en)	88.13432835820896	59.298209334395665	83.31769058643586
MTEB AmazonPolarityClassification	91.526375	88.16327709705504	91.51095801287843
MTEB AmazonReviewsClassification (en)	47.856	-	45.41490917650942
MTEB Banking77Classification	78.51298701298701	-	77.49041754069235

Retrieval Tasks

Dataset	MAP@1	MAP@10	MAP@100	MAP@1000	MRR@1	MRR@10	MRR@100	MRR@1000	NDCG@1	NDCG@10	NDCG@100	NDCG@1000	Precision@1	Precision@10	Precision@100	Precision@1000	Recall@1	Recall@10	Recall@100	Recall@1000
MTEB ArguAna	31.223	47.947	48.742000000000004	48.745	32.432	48.4	49.202	49.205	31.223	57.045	60.175	60.233000000000004	31.223	8.599	0.991	0.1	31.223	85.989	99.075	99.502
MTEB CQADupstackAndroidRetrieval	35.803000000000004	48.848	50.5	50.602999999999994	44.635000000000005	55.593	56.169999999999995	56.19499999999999	44.635000000000005	55.899	60.958	62.302	44.635000000000005	10.786999999999999	1.6580000000000001	0.213	35.803000000000004	68.657	89.77199999999999	97.67
MTEB CQADupstackEnglishRetrieval	33.706	44.896	46.299	46.44	41.592	50.529	51.22	51.258	41.592	50.77199999999999	55.383	57.288	41.592	9.516	1.541	0.2	33.706	61.353	80.182	91.896
MTEB CQADupstackGamingRetrieval	44.424	57.169000000000004	58.202	58.242000000000004	50.470000000000006	60.489000000000004	61.096	61.112	50.470000000000006	63.071999999999996	66.964	67.659	50.470000000000006	10.019	1.29	0.13899999999999998	44.424	77.02	93.738	98.451
MTEB CQADupstackGisRetrieval	26.294	34.503	35.641	35.724000000000004	28.362	36.53	37.541000000000004	37.602000000000004	28.362	39.513999999999996	44.815	46.839	28.362	6.101999999999999	0.9129999999999999	0.11399999999999999	26.294	53.098	76.877	91.834
MTEB CQADupstackMathematicaRetrieval	16.407	25.185999999999996	26.533	26.657999999999998	20.522000000000002	29.522	30.644	30.713	20.522000000000002	30.656	36.864999999999995	39.675	20.522000000000002	5.795999999999999	1.027	-	16.407	-	-	-

Clustering Tasks

Dataset	V - Measure
MTEB ArxivClusteringP2P	43.1621946393635
MTEB ArxivClusteringS2S	32.56417132407894
MTEB BiorxivClusteringP2P	37.61848554098577
MTEB BiorxivClusteringS2S	31.32623280148178

Reranking Task

Dataset	MAP	MRR
MTEB AskUbuntuDupQuestions	64.29539304390207	76.44484017060196

STS Task

Dataset	Cosine Similarity Spearman
MTEB BIOSSES	84.38746499431112

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