Datasets:
text_id stringlengths 6 6 | text stringlengths 20 134 | split stringclasses 1
value | category stringclasses 10
values | origin stringclasses 3
values | source_license stringclasses 2
values | source_url stringclasses 3
values |
|---|---|---|---|---|---|---|
t00000 | Outono e inverno, para o velho tempo do inferno. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00001 | Celso Lopes de Freitas. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00002 | A gente se irrita quando eles falam do passado com cara de sofrimento. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00003 | Mas então percebi que ela não poderia ter sido. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00004 | Tá tudo arrumado pra festa do casamento, só faltam os convidados. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00005 | Você acha que a gente pode fazer caminhada no Parque do Pão de Açúcar essa semana? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00006 | O estado governante não gerencia. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00007 | Vou te contar, a gente tá com um pé na lama aqui no bairro. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00008 | O carro da Ana quebrou no meio do caminho e ela ficou chorando, mas a gente vai consertar! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00009 | Tem uns pratos que são fáceis de fazer, como arroz com frango, né? | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00010 | Os alunos tingem suas cores. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00011 | A gente tá com o rosto colorido e a cara de quem acaba de ganhar um prêmio! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00012 | O João não gosta de atrasos, mas sempre acontece. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00013 | Tá tudo bem com o plano, a gente só precisa seguir em frente. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00014 | Se você não ganhou dinheiro por quarenta anos, nos anos sessenta não fará nada. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00015 | Ontem a gente saiu e tá com frio, a gente precisa de um cobertor. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00016 | A gente não precisa mais usar o ônibus, tá economizando, né? | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00017 | Você acha que a gente pode levar crianças pro parque esse fim de semana? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00018 | Insira seu cartão para se identificar. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00019 | A gente tem que ficar ligado, a gente nunca sabe o que pode acontecer. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00020 | O cachorro preto está pulando em alguma coisa no quintal. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00021 | Você acha que o hotel tem internet gratuita né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00022 | Gritando alguma coisa, não me mate. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00023 | Talvez nós somos todos assim? o menino meditou. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00024 | não digo que se deixasse rolar pelo chão, convulsa. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00025 | O prazo para entregar o trabalho termina no dia vinte e dois de fevereiro, numa terça-feira. | train | data | template | CC-BY-4.0 | |
t00026 | Você já comeu churrasco de domingo antes? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00027 | O churrasco de domingo tá ficando pra semana que vem. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00028 | A gente já tá cansado de tantas paradas no caminho. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00029 | Três homens em um rio sujo. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00031 | A gente quer que o trânsito seja mais tranquilo, né? | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00032 | Você leu aquele livro da Ana que eu recomendei? | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00033 | uma criança pequena em uma camisa rosa correndo através de um campo florido. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00034 | O ônibus passa às quatro horas e quinze da noite. | train | hora | template | CC-BY-4.0 | |
t00035 | Nós viemos de Alcalali. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00037 | Por favor, chame o senhor Carlos Santos na recepção. | train | nome | template | CC-BY-4.0 | |
t00038 | A gente estava no meu apartamento e resolvi fazer um bolo de chocolate pra surpreender a minha namorada. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00039 | A gente já viu o jogo do Cuiabá? Foi muito disputado, cara. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00040 | O dia foi esplêndido; Um verdadeiro domingo de Páscoa. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00041 | Você pode me fazer um pix de oitenta e quatro reais? | train | dinheiro | template | CC-BY-4.0 | |
t00042 | O goleiro do Zé Carlos errou uma bola na marca do pênalti e a galera ficou pra lá de braba. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00044 | Você tá me salvando, cara, essa ajuda foi show de bola! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00045 | Eu quase me esqueci depois de comer. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00046 | A gente tá com fome e essa cidade grande tem de tudo pra satisfazer o apetite! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00047 | A gente chegou cedo pra ver o show da Luana no festival do verão. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00048 | Você vai pra onde essa viagem de férias tá rolando né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00049 | Tem muitos carros, mas a gente não consegue se orientar. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00050 | A gente não pode ignorar o que o médico disse, ele tá certo né! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00051 | Tá difícil achar o tempero certo pra esse prato de carne. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00052 | Tem uns celulares que a gente nem sabe o que tá dentro, só sabe que é bom. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00053 | Se uma atividade específica é realizada que promove a interculturalidade: dez pontos. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00055 | Mãe e filho se divertindo comendo à beira do lago. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00056 | O João tá tentando ligar o Wi-Fi no quarto sem sucesso. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00058 | Você tá atrasado pra chegar no escritório essa semana né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00059 | A gente chegou cedo e o recepcionista foi super gentil. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00060 | Aurina Gomes da Silva Moreira. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00062 | Meu primo foi pra Bahia pra aprender a fazer o cajá na roça. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00063 | Na faculdade, a gente tinha mais liberdade, mas também mais responsabilidades. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00064 | Cozinhar em casa é tão bom que a gente até esquece de tudo! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00065 | A gente já viu tanta diversão na quadra da escola! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00066 | Você sacrificaria a vida de um humano se isso salvasse muitas outras vidas? | train | pergunta | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00067 | Como a gente evita que o trabalho invada o tempo de lazer? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00069 | Você viu a gente no hospital ontem? Tá tudo bem com o seu irmão? A gente tem que ficar de olho nele né. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00070 | Você acha que a gente tá atrasado pra chegar na praia? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00071 | A gente vai ter um bolo gigante pra comemorar o casamento. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00072 | Meu pai não fuma nem bebe. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00073 | Não na mesma hora que eu encontrei. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00074 | Estágio Universitário de Experiência de Trabalho Juvenil. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00075 | O João falou que o verão vai durar até setembro. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00076 | Tinta cinza se estendia por quilômetros ao redor. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00077 | Essa mudança tá me deixando nervoso, porque a gente vai ter que arrumar tudo de novo. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00078 | Agora será com esperança. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00079 | A gente ia pagar a conta e o app do banco não carregava! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00080 | Fizemos o check-in cedo, mas o guia chegou atrasado e a gente ficou esperando horas. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00082 | A gente pode pedir pra alguém que tá com a gente ir no bar né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00083 | O Tratado de Tordesilhas foi firmado por Portugal e a Coroa de Castela. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00084 | Você acha que o novo corte do Lucas fica legal pra ele né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00086 | A empresa fez uma reunião com os clientes e falou sobre melhorias no atendimento, a gente aprovou tudo. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00087 | Pra mim, só tem cinema legal se tiver um final surpreendente. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00088 | O correio daqui tá muito lento, parece que tá parado. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00089 | "Uma prática de infiéis?" ele disse para si mesmo. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00090 | Você recebeu uma mensagem de áudio no meio do expediente e não sabia se era importante ou não. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00091 | Cem freiras invadiram o Vaticano. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00092 | Eu vou dormir na pequena cama no quarto. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00094 | A consulta foi marcada para as quinze horas e meia. | train | hora | template | CC-BY-4.0 | |
t00095 | Meu avô falou que a eleição é a forma mais importante de participar da vida do país, e a gente concorda. | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00096 | A pandemia escancarou a irracionalidade do sistema capitalista. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00097 | André, você viu onde eu deixei a chave do carro? | train | nome | template | CC-BY-4.0 | |
t00098 | O funcionário do cartório foi super rude, foi uma experiência péssima, cara! | train | exclamacao | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00099 | Eu pensei que você ia me ensinar algumas das coisas que você sabe. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00100 | Você já fez bolo com frutas do seu quintal pra deixar mais fresco né? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00101 | Paguei dezoito centavos no cafezinho da padaria. | train | dinheiro | template | CC-BY-4.0 | |
t00102 | A gente precisa de check-up, não adianta ignorar os sinais. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00103 | Você acha que tá cedo demais pra ir pra praia hoje? | train | pergunta | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00104 | Tá me falando que jogar é uma forma de se desconectar da realidade, né? | train | narrativa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00105 | Caindo no chão com uma bengala. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00106 | Isso salvou sua vida. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00107 | A gente tá tentando entender o que está errado com o frete, não dá pra entender. | train | conversa | llm:Qwen/Qwen3-8B | CC-BY-4.0 | https://huggingface.co/Qwen/Qwen3-8B |
t00108 | Os limites eram bem definidos. | train | leitura | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
t00109 | Por que não aqui mesmo? | train | pergunta | common_voice:sentence-collector | CC0-1.0 | https://github.com/common-voice/common-voice/blob/main/server/data/pt/sentence-collector.txt |
SemFence PT-BR Sentences
Um banco aberto de 10,675 frases em português do Brasil, feito para ser lido em voz alta: conversa do dia a dia, perguntas, exclamações, narrativas, e números, datas, horários, dinheiro e nomes escritos por extenso. An open bank of 10,675 Brazilian Portuguese sentences made to be read aloud: everyday conversation, questions, exclamations, narrative, and numbers, dates, times, money and names written out in words.
Created by SemFence · researcher: Gustavo Toledo. These sentences are the text of semfence/ptbr-synthetic-speech, a synthetic PT-BR speech dataset, and can be used on their own for TTS, ASR, language-model or NLP work.
| Sentences | 10,675 |
| Language | Portuguese (Brazil) |
| Length | 2–28 words, median 11 |
| Splits | train 9,937 · validation 427 · test 311 |
| License | CC BY 4.0, credit required (see "License and attribution") |
Usage
from datasets import load_dataset
ds = load_dataset("semfence/ptbr-sentences")
print(ds["train"][0])
Where the sentences come from
| Origin | Sentences |
|---|---|
| llm | 5,355 |
| common_voice | 4,200 |
| template | 1,120 |
| Origin | What it is | License of the text at its source |
|---|---|---|
common_voice |
Brazilian Portuguese sentences from the Mozilla Common Voice sentence collector, filtered as described below | CC0 1.0 |
llm |
Conversational sentences written for this project by Qwen3-8B (Apache-2.0), one batch per topic and style (statement, question, exclamation, narrative) over 70 everyday topics. The prompts are in the speech dataset repository, under generation/ |
CC BY 4.0 (SemFence) |
template |
Sentences built from templates with num2words: money, dates, times, quantities, ages, names and questions |
CC BY 4.0 (SemFence) |
Every row records its origin, source_license and source_url.
How it was filtered
- Between 20 and 220 characters; no digits or symbols (numbers are written in words, so no text normalizer is needed to read them).
- Exact duplicates removed (ignoring case and punctuation).
- Constructions typical of European Portuguese were dropped (for example enclitic pronouns such as "-lhe" or "-vos" and words such as "autocarro" or "pequeno-almoço").
- Sentences without final punctuation received one: "?" for questions, "!" for exclamations, "." otherwise.
- Split by sentence into train, validation and test (about 93%, 4% and 3%).
Columns
| Column | Description |
|---|---|
text_id |
Sentence ID (t00000…). The same ID is used in the speech dataset |
text |
The sentence |
split |
train, validation or test |
category |
Kind of sentence. Common Voice: leitura (statements), pergunta and exclamacao (sorted by the final punctuation). LLM: conversa, pergunta, exclamacao, narrativa. Templates: dinheiro, data, hora, numero, nome, pergunta |
origin |
Where the sentence came from |
source_license, source_url |
License of the text at its source, and a link |
| Category | Sentences |
|---|---|
| leitura | 3,871 |
| pergunta | 1,748 |
| exclamacao | 1,397 |
| conversa | 1,351 |
| narrativa | 1,308 |
| data | 220 |
| hora | 220 |
| dinheiro | 220 |
| numero | 220 |
| nome | 120 |
Limitations
- The
llmsentences were not reviewed by a person. They passed only automatic filters, so some may sound odd, repeat a pattern (many start with "A gente…") or be factually silly. - The
common_voicesentences are mostly read-style text and many are short. A few may still read as European Portuguese or contain unusual spellings, because the filter is automatic. - The topics are everyday ones. There is little specialized, technical or regional vocabulary.
License and attribution
This dataset is released under CC BY 4.0. You may use, share and adapt it, including commercially, as long as you give credit:
SemFence PT-BR Sentences, by SemFence (Gustavo Toledo) — https://huggingface.co/datasets/semfence/ptbr-sentences
The Common Voice sentences are CC0 at their source and stay CC0 there. This compilation, its annotations and the llm and template sentences are CC BY 4.0, so the credit above is required when you use the dataset.
Citation
@misc{semfence2026ptbrsentences,
author = {{SemFence} and Toledo, Gustavo},
title = {SemFence PT-BR Sentences: an open Brazilian Portuguese sentence bank},
year = {2026},
howpublished = {Hugging Face Datasets},
url = {https://huggingface.co/datasets/semfence/ptbr-sentences}
}
Acknowledgements
Mozilla Common Voice (sentence collector, CC0); Qwen3-8B by Alibaba Qwen (Apache-2.0), used to write the conversational sentences; num2words (LGPL), used by the templates. The recipe follows the open tiny-voca dataset by Muhammed Nazeem.
- Downloads last month
- 65