CoolFace
14 shown

datasets

Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.

modalityAllTextImageAudioTabularVideo
taskAllText generationQuestion answeringImage classificationSpeech
sizeAllUnder 1K1K – 10K100K – 1M1M – 10M
formatAllParquetJSONCSVWebDataset
sortTrendingMost likedMost usedRecently updatedNewest
Clear all
01litagin /moe-speechgated Recomendation of OOPPEENN's 56697375616C4E6F76656C5F44617461736574 I recommend OOPPEENN/56697375616C4E6F76656C5F44617461736574 for Japanese voice corpus, which is: Similar speech domain to this one (Japanese anime-style speech from Japanese Visual Novel), but Huge amounts of audio compared to this dataset (600 hours for this, 10,000 hours for Galgame_Dataset!) This dataset contains about 50 games, and Galgame_Dataset contains more than 500 games! Contains true transcripts of each… See the full description on the dataset page: https://huggingface.co/datasets/litagin/moe-speech.audiotext-to-speech100K<n<1M280 likes667 downloads5mo agoHugging Face02Uni-MoE /College-Entrance-English-Examination-Listening-PartIf our dataset is useful for your research, please cite our work: @article{li2025uni, title={Uni-moe: Scaling unified multimodal llms with mixture of experts}, author={Li, Yunxin and Jiang, Shenyuan and Hu, Baotian and Wang, Longyue and Zhong, Wanqi and Luo, Wenhan and Ma, Lin and Zhang, Min}, journal={IEEE Transactions on Pattern Analysis and Machine Intelligence}, year={2025}, publisher={IEEE} } audion<1K1 likes128 downloads1y agoHugging Face03NickWeng /tat-moegated TAT-MOE TAT (Taiwanese Across Taiwan) MOE 台文語音語料庫 -- speakers reading aloud example sentences from the Ministry of Education's official Taiwanese dictionary (臺灣台語常用詞 辭典), recorded simultaneously on 6 different microphones per sentence. ⚠️ Known incompleteness -- this is NOT the full official TAT-MOE corpus. The official release covers 328 train / 58 eval / 54 test speakers (86,072 / 16,357 / 15,962 sentences). What is actually available in the source COS bucket -- and therefore… See the full description on the dataset page: https://huggingface.co/datasets/NickWeng/tat-moe.audio1 likes115 downloads14d agoHugging Face04Backup-bdg /moe-training Xoron-Dev Multimodal MoE Dataset This dataset is a high-scale, unified multimodal collection specifically engineered for training Mixture of Experts (MoE) models. It integrates text, audio, image, and video data into a single, cohesive training pipeline designed to foster cross-modal reasoning, creative generation, and agentic behavior. 🚀 Capabilities By utilizing this dataset, models can be trained for: Vision-Language: Image generation, high-fidelity editing… See the full description on the dataset page: https://huggingface.co/datasets/Backup-bdg/moe-training.image100K<n<1M2 likes64 downloads7mo agoHugging Face05alexachang /tat-moe-subsetaudio10K<n<100K0 likes62 downloads1y agoHugging Face06ayousanz /moe-speech-plusgated MoeSpeechPlus MoeSpeechのデータセットに以下のデータ処理を行ったデータセット 追加情報 anime-whisperによる文字お越し parakeetによる文字お越し speechMOS (UTMOS v2 (T05)を使用) 音声の長さ microsoft/deberta-v3-largeによる感情判定 litagin/anime_speech_emotion_classificationによる感情判定 Qwen/Qwen2-Audio-7B-Instructによる感情判定 { "parakeet_jp_transcription": "昨夜からずっと気配を探られていたか。", "anime_whisper_transcription": "昨夜からずっと気配を探られていたか…", "duration": 3.722018140589569, "speechMOS": 2.294760227203369, "DeBERTa_Sentiment": { "悲しみ": 0.12533913552761078… See the full description on the dataset page: https://huggingface.co/datasets/ayousanz/moe-speech-plus.audiotext-to-speech100K<n<1M8 likes37 downloads1y agoHugging Face07ayousanz /moe-speech-20speakers-ljspeechgated MoeSpeech 20 Speakers - LJSpeech Format moe-speech-plus-ljspeechから発話数上位20話者を抽出したデータセットです。 Piper TTSなどのTTSモデル学習に最適化されています。 Dataset Summary 項目 値 話者数 20 総発話数 60,233 フォーマット LJSpeech サンプルレート 22050 Hz 言語 日本語 総音声時間 約90.6時間 Speaker Statistics Speaker ID Utterances Internal ID 940de876 4,675 0 2cf01874 4,632 1 bbd90363 3,747 2 1a5a3db8 3,295 3 4e2f4ba6 3,268 4 cc948b89 3,043 5 ad28b91b 3,033 6 ee093a4f 2,957 7… See the full description on the dataset page: https://huggingface.co/datasets/ayousanz/moe-speech-20speakers-ljspeech.audiotext-to-speech10K<n<100K1 likes21 downloads9mo agoHugging Face08NickWeng /moe-dictgated MOE Taiwanese Dictionary (台語字典) Word/phrase-level dictionary entry recordings, likely sourced from the Ministry of Education's official Taiwanese dictionary (臺灣台語常用詞辭典) -- each entry is a single word or short phrase (not a full sentence), with Han-lo text, a Mandarin gloss, and Tai-lo romanization. Dataset Structure id: numeric dictionary-entry identifier (e.g. 45549), already globally unique. audio: audio clip (16kHz, mono, 24-bit PCM WAV (majority)). text:… See the full description on the dataset page: https://huggingface.co/datasets/NickWeng/moe-dict.audio10K<n<100K0 likes20 downloads19d agoHugging Face09ayousanz /moe-speech-5speakers-ljspeechgated moe-speech-5speakers-ljspeech Japanese multi-speaker TTS dataset in LJSpeech format. Overview Speakers: 5 Total utterances: 19,617 Format: LJSpeech (wavs/ + metadata.csv) Sample rate: 22050 Hz Language: Japanese Speaker Statistics Speaker ID Utterances 940de876 4,675 2cf01874 4,632 bbd90363 3,747 1a5a3db8 3,295 4e2f4ba6 3,268 Files metadata.csv - Full metadata (id|speaker|text) metadata_{speaker_id}.csv - Per-speaker… See the full description on the dataset page: https://huggingface.co/datasets/ayousanz/moe-speech-5speakers-ljspeech.audio10K<n<100K0 likes8 downloads10mo agoHugging Face10formospeech /tat_moe_cmgated TEST Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents nan_tw Taigi 12.89 12,578 151,234 3.69 3.26 0 0 Total - 12.89 12,578 151,234 3.69 3.26 0 0 EVAL Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents nan_tw Taigi 11.68 10,766 134,941 3.91 3.21 0 0 Total - 11.68 10,766 134,941 3.91 3.21 0 0 TRAIN Subset lang_name hours n_utts… See the full description on the dataset page: https://huggingface.co/datasets/formospeech/tat_moe_cm.audio10K<n<100K0 likes8 downloads1mo agoHugging Face11formospeech /hakkadict_moe_examplegated TRAIN Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents hak_dp Hakka_Dapu 0.00 0 0 0.00 0.00 14,455 282,665 hak_hl Hakka_Hailu 0.00 0 0 0.00 0.00 14,996 294,876 hak_nsx Hakka_NanSixian 0.00 0 0 0.00 0.00 14,776 291,321 hak_rp Hakka_Raoping 0.00 0 0 0.00 0.00 14,881 293,945 hak_sx Hakka_Sixian 0.00 0 0 0.00 0.00 15,010 299,058 hak_za Hakka_Zhaoan 0.00 0 0 0.00 0.00 12,530 246,820 Total - 0.00 0 0 0.00 0.00 86,648 1… See the full description on the dataset page: https://huggingface.co/datasets/formospeech/hakkadict_moe_example.audio10K<n<100K1 likes6 downloads3mo agoHugging Face12formospeech /tat_moe_mcgated TEST Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents nan_tw Taigi 12.86 12,582 151,247 3.68 3.27 0 0 Total - 12.86 12,582 151,247 3.68 3.27 0 0 EVAL Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents nan_tw Taigi 11.74 10,769 134,953 3.92 3.19 0 0 Total - 11.74 10,769 134,953 3.92 3.19 0 0 TRAIN Subset lang_name hours n_utts… See the full description on the dataset page: https://huggingface.co/datasets/formospeech/tat_moe_mc.audio10K<n<100K0 likes5 downloads1mo agoHugging Face13formospeech /hakkadict_moe_wordgated TRAIN Subset lang_name hours n_utts n_chars_in_utts secs/utt chars/sec n_sents n_chars_in_sents hak_dp Hakka_Dapu 7.05 11,603 25,919 2.19 1.02 0 0 hak_hl Hakka_Hailu 4.56 11,403 25,468 1.44 1.55 0 0 hak_nsx Hakka_NanSixian 4.81 11,676 26,283 1.48 1.52 0 0 hak_rp Hakka_Raoping 3.68 11,391 25,495 1.16 1.93 0 0 hak_sx Hakka_Sixian 3.86 11,393 25,519 1.22 1.84 0 0 hak_za Hakka_Zhaoan 3.31 9,946 21,265 1.20 1.78 0 0 Total - 27.27 67,412 149,949 1.46 1.53 0 0… See the full description on the dataset page: https://huggingface.co/datasets/formospeech/hakkadict_moe_word.audio10K<n<100K0 likes4 downloads2mo agoHugging Face14ayousanz /moe-speech-plus-ljspeechgated MoeSpeechPlus-ljspeech MoeSpeechPlusのデータセットのデータフォーマットをljspechに沿って変換したもの huggingface-cli download --repo-type dataset ayousanz/moe-speech-plus-ljspeech --local-dir moe-speech-plus-ljspeech MoeSpeech 日本語はこちら このデータセットは、著作権法第三十条の四の情報解析（機械学習等）の目的でのみ使用が許可されています。それ以外の用途での使用はライセンスにより禁止されています。 This dataset is only permitted for use under Article 30-4 of the Copyright Law of Japan for data analysis (such as machine learning) purposes. Any use for purposes other than those… See the full description on the dataset page: https://huggingface.co/datasets/ayousanz/moe-speech-plus-ljspeech.audiotext-to-speech100K<n<1M1 likes2 downloads1y agoHugging Face

Listings come live from the Hugging Face Hub API. CoolFace does not host these files.