Multilingual-E5-large
Comprehensive information on the functionality and usage of the Multilingual-E5-large model
API Reference: Embeddings API
Model Reference: Multilingual-E5-large
Paper: Text Embeddings by Weakly-Supervised Contrastive Pre-training
Layers | Embedding Dimension | Recommended Sequence Length |
---|---|---|
24 | 1024 | 512 |
Suitable Score Functions
- cosine-similarity
Supported Languages
This model is initialized from xlm-roberta-large and continually trained on a mixture of multilingual datasets. It supports 100 languages from xlm-roberta, but low-resource languages may see performance degradation.
Working with instruction
Each input text should start with "query" or "passage", even for non-English texts. For tasks other than retrieval, you can simply use the query
instruction.
This can be passed the instruction
parameter in the request. By default we calculate the embeddings the instruction query
.
Examples
Calculate Sentence similarities
Similarities
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
import requests from sklearn.metrics.pairwise import cosine_similarity headers = {'Content-Type': 'application/json', 'Authorization': 'Bearer {EMBAAS_API_KEY}'} science_sentences = [ "Parton energy loss in QCD matter", "The Chiral Phase Transition in Dissipative Dynamics" ] instruction = "query" data = { 'texts': science_sentences, 'model': 'multilingual-e5-large', 'instruction': instruction } response = requests.post("https://api.embaas.io/v1/embeddings/", json=data, headers=headers) embeddings = response.json()["data"] similarities = cosine_similarity([embeddings[0]["embedding"]], [embeddings[1]["embedding"]]) print(similarities)
Information Retrieval
Retrieval
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
import requests import numpy as np from sklearn.metrics.pairwise import cosine_similarity headers = {'Content-Type': 'application/json', 'Authorization': 'Bearer {EMBAAS_API_KEY}'} def get_embeddings(texts, model, instruction=None): data = {'texts': texts, 'model': model} if instruction: data['instruction'] = instruction response = requests.post("https://api.embaas.io/v1/embeddings/", json=data, headers=headers) embeddings = [entry['embedding'] for entry in response.json()['data']] return np.array(embeddings) query_instruction = "query" query_text = "where is the food stored in a yam plant" corpus_instruction = "passage" corpus_texts = [ "Capitalism has been dominant in the Western world since the end of feudalism, but most feel[who?] that the term 'mixed economies' more precisely describes most contemporary economies, due to their containing both private-owned and state-owned enterprises. In capitalism, prices determine the demand-supply scale. For example, higher demand for certain goods and services lead to higher prices and lower demand for certain goods lead to lower prices.", "The disparate impact theory is especially controversial under the Fair Housing Act because the Act regulates many activities relating to housing, insurance, and mortgage loans—and some scholars have argued that the theory's use under the Fair Housing Act, combined with extensions of the Community Reinvestment Act, contributed to rise of sub-prime lending and the crash of the U.S. housing market and ensuing global economic recession", "Disparate impact in United States labor law refers to practices in employment, housing, and other areas that adversely affect one group of people of a protected characteristic more than another, even though rules applied by employers or landlords are formally neutral. Although the protected classes vary by statute, most federal civil rights laws protect based on race, color, religion, national origin, and sex as protected traits, and some laws include disability status and other traits as well." ] model_name = "multilingual-e5-large" query_embeddings = get_embeddings([query_text], model_name, query_instruction) corpus_embeddings = get_embeddings(corpus_texts, model_name, corpus_instruction) similarities = cosine_similarity(query_embeddings, corpus_embeddings) retrieved_doc_id = np.argmax(similarities) print(corpus_texts[retrieved_doc_id])