MT0-XL Open-Source Multilingual Text Generation Model - Free Support for Coreference Resolution and Natural Language Inference

Mt0 Xl

Developed by bigscience

MT0-XL is a multilingual text generation model that supports multiple languages and tasks, including coreference resolution and natural language inference.

Large Language Model

Transformers

Supports Multiple LanguagesOpen Source License:Apache-2.0 #Multilingual Coreference Resolution #Cross-lingual Natural Language Inference #Low-resource Language Support

Downloads 2,807

Release Time : 10/27/2022

Model Overview

MT0-XL is a multilingual model based on the Transformer architecture, suitable for various natural language processing tasks such as coreference resolution and natural language inference.

Model Features

Multilingual Support

Supports over 100 languages, suitable for natural language processing tasks worldwide.

Multi-task Processing

Capable of performing various natural language processing tasks, including coreference resolution and natural language inference.

High Performance

Outstanding performance in multiple benchmarks, especially in coreference resolution tasks.

Model Capabilities

Text generation

Coreference resolution

Natural language inference

Use Cases

Natural Language Processing

Coreference Resolution

Used to resolve reference issues in text, improving the accuracy of text understanding.

Achieved an accuracy of 52.49% on the Winogrande XL dataset.

Natural Language Inference

Used to determine the logical relationship between two sentences (such as entailment, contradiction, or neutrality).

Achieved accuracies of 38.2% (r1), 34.8% (r2), and 39% (r3) on the ANLI dataset.

🚀 MT0-XL Model README

This document provides detailed information about the mt0-xl model, including its datasets, supported languages, license, and performance metrics.

📦 Datasets

The model is trained on the following datasets:

bigscience/xP3
mc4

📄 License

The model is released under the apache-2.0 license.

🌐 Supported Languages

The model supports a wide range of languages, including:

af, am, ar, az, be, bg, bn, ca, ceb, co, cs, cy, da, de, el, en, eo, es, et, eu, fa, fi, fil, fr, fy, ga, gd, gl, gu, ha, haw, hi, hmn, ht, hu, hy, ig, is, it, iw, ja, jv, ka, kk, km, kn, ko, ku, ky, la, lb, lo, lt, lv, mg, mi, mk, ml, mn, mr, ms, mt, my, ne, nl, 'no', ny, pa, pl, ps, pt, ro, ru, sd, si, sk, sl, sm, sn, so, sq, sr, st, su, sv, sw, ta, te, tg, th, tr, uk, und, ur, uz, vi, xh, yi, yo, zh, zu

🛠️ Pipeline Tag

The model's pipeline tag is text2text-generation.

🧪 Widget Examples

The following are some example texts and their corresponding example titles for the model's widget:

Example Title	Text
zh-en sentiment	一个传奇的开端，一个不灭的神话，这不仅仅是一部电影，而是作为一个走进新时代的标签，永远彪炳史册。Would you rate the previous review as positive, neutral or negative?
zh-zh sentiment	一个传奇的开端，一个不灭的神话，这不仅仅是一部电影，而是作为一个走进新时代的标签，永远彪炳史册。你认为这句话的立场是赞扬、中立还是批评？
vi-en query	Suggest at least five related search terms to "Mạng neural nhân tạo".
fr-fr query	Proposez au moins cinq mots clés concernant «Réseau de neurones artificiels».
te-en qa	Explain in a sentence in Telugu what is backpropagation in neural networks.
en-en qa	Why is the sky blue?
es-en fable	Write a fairy tale about a troll saving a princess from a dangerous dragon. The fairy tale is a masterpiece that has achieved praise worldwide and its moral is "Heroes Come in All Shapes and Sizes". Story (in Spanish):
hi-en fable	Write a fable about wood elves living in a forest that is suddenly invaded by ogres. The fable is a masterpiece that has achieved praise worldwide and its moral is "Violence is the last refuge of the incompetent". Fable (in Hindi):

📊 Model Performance

Model Name: mt0-xl

The following table shows the model's performance on various tasks and datasets:

Task Type	Dataset Type	Dataset Name	Config	Split	Revision	Metric Type	Value
Coreference resolution	winogrande	Winogrande XL (xl)	xl	validation	a80f460359d1e9a67c006011c94de42a8759430c	Accuracy	52.49
Coreference resolution	Muennighoff/xwinograd	XWinograd (en)	en	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	61.89
Coreference resolution	Muennighoff/xwinograd	XWinograd (fr)	fr	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	59.04
Coreference resolution	Muennighoff/xwinograd	XWinograd (jp)	jp	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	60.27
Coreference resolution	Muennighoff/xwinograd	XWinograd (pt)	pt	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	66.16
Coreference resolution	Muennighoff/xwinograd	XWinograd (ru)	ru	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	59.05
Coreference resolution	Muennighoff/xwinograd	XWinograd (zh)	zh	test	9dd5ea5505fad86b7bedad667955577815300cee	Accuracy	62.9
Natural language inference	anli	ANLI (r1)	r1	validation	9dbd830a06fea8b1c49d6e5ef2004a08d9f45094	Accuracy	38.2
Natural language inference	anli	ANLI (r2)	r2	validation	9dbd830a06fea8b1c49d6e5ef2004a08d9f45094	Accuracy	34.8
Natural language inference	anli	ANLI (r3)	r3	validation	9dbd830a06fea8b1c49d6e5ef2004a08d9f45094	Accuracy	39
Natural language inference	super_glue	SuperGLUE (cb)	cb	validation	9e12063561e7e6c79099feb6d5a493142584e9e2	Accuracy	85.71
Natural language inference	super_glue	SuperGLUE (rte)	rte	validation	9e12063561e7e6c79099feb6d5a493142584e9e2	Accuracy	78.7
Natural language inference	xnli	XNLI (ar)	ar	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	51.85
Natural language inference	xnli	XNLI (bg)	bg	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	54.18
Natural language inference	xnli	XNLI (de)	de	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	54.78
Natural language inference	xnli	XNLI (el)	el	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	53.78
Natural language inference	xnli	XNLI (en)	en	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	56.83
Natural language inference	xnli	XNLI (es)	es	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	54.78
Natural language inference	xnli	XNLI (fr)	fr	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	54.22
Natural language inference	xnli	XNLI (hi)	hi	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	50.24
Natural language inference	xnli	XNLI (ru)	ru	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	53.09
Natural language inference	xnli	XNLI (sw)	sw	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	49.6
Natural language inference	xnli	XNLI (th)	th	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	52.13
Natural language inference	xnli	XNLI (tr)	tr	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	50.56
Natural language inference	xnli	XNLI (ur)	ur	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	47.91
Natural language inference	xnli	XNLI (vi)	vi	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	53.21
Natural language inference	xnli	XNLI (zh)	zh	validation	a5a45e4ff92d5d3f34de70aaf4b72c3bdf9f7f16	Accuracy	50.64
Program synthesis	openai_humaneval	HumanEval	None	test	e8dc562f5de170c54b5481011dd9f4fa04845771	Pass@1	0
Program synthesis	openai_humaneval	HumanEval	None	test	e8dc562f5de170c54b5481011dd9f4fa04845771	Pass@10	0
Program synthesis	openai_humaneval	HumanEval	None	test	e8dc562f5de170c54b5481011dd9f4fa04845771	Pass@100	0
Sentence completion	story_cloze	StoryCloze (2016)	'2016'	validation	e724c6f8cdf7c7a2fb229d862226e15b023ee4db	Accuracy	79.1
Sentence completion	super_glue	SuperGLUE (copa)	copa	validation	9e12063561e7e6c79099feb6d5a493142584e9e2	Accuracy	72
Sentence completion	xcopa	XCOPA (et)	et	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	70
Sentence completion	xcopa	XCOPA (ht)	ht	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	66
Sentence completion	xcopa	XCOPA (id)	id	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	71
Sentence completion	xcopa	XCOPA (it)	it	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	70
Sentence completion	xcopa	XCOPA (qu)	qu	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	56
Sentence completion	xcopa	XCOPA (sw)	sw	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	53
Sentence completion	xcopa	XCOPA (ta)	ta	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	64
Sentence completion	xcopa	XCOPA (th)	th	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	60
Sentence completion	xcopa	XCOPA (tr)	tr	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	58
Sentence completion	xcopa	XCOPA (vi)	vi	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	68
Sentence completion	xcopa	XCOPA (zh)	zh	validation	37f73c60fb123111fa5af5f9b705d0b3747fd187	Accuracy	65
Sentence completion	Muennighoff/xstory_cloze	XStoryCloze (ar)	ar	validation	8bb76e594b68147f1a430e86829d07189622b90d	Accuracy	70.09
Sentence completion	Muennighoff/xstory_cloze	XStoryCloze (es)	es	validation	8bb76e594b68147f1a430e86829d07189622b90d	Accuracy	77.17

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