FERNAN89/mTEDx-ptbr

Multilingual TEDx (Portuguese speech and transcripts) NOTE: This dataset contains only the Portuguese portion of the mTEDx dataset, already processed and segmented into parts. Multilingual TEDx (mTEDx) is a multilingual speech recognition and translation corpus to facilitate the training of ASR and SLT models in additional languages. The corpus comprises audio recordings and transcripts from TEDx Talks in 8 languages (Spanish, French, Portuguese, Italian, Russian, Greek, Arabic… See the full description on the dataset page: https://huggingface.co/datasets/FERNAN89/mTEDx-ptbr.