- https://t.me/speech_recognition_ru - группа "Распознавание речи"
- https://t.me/speech_recognition - группа на английском языке
- https://t.me/speechtech - канал новостей
- https://t.me/betterdatacommunity/15 - сообщество speech в datacommunity
- https://t.me/voicestuff https://t.me/voice_stuff_chat - Frappucсino's space
- https://t.me/TeraSpace https://t.me/teraspace_chat - Tera's space
- https://github.com/markovka17/dla
- https://github.com/yandexdataschool/speech_course
- https://github.com/severilov/DL-Audio-Course
- https://huggingface.co/learn/audio-course/ru/chapter0/introduction - поиграться со звуковыми моделями HF
- https://www.youtube.com/playlist?list=PLYG3WHDP5CWVRxLjXZbllqIQTWY_QjKmz - Deep Learning for Audio
- https://github.com/salute-developers/golos
- https://github.com/snakers4/open_stt
- https://github.com/GeorgeFedoseev/DeepSpeech
- https://github.com/sovaai/sova-dataset
- https://www.openslr.org/96/ - Russian Librispeech
- https://commonvoice.mozilla.org/ru/datasets - MCV
- https://www.caito.de/2019/01/03/the-m-ailabs-speech-dataset/ - M-AILabs dataset (from Librivox)
- https://ruslan-corpus.github.io/
- https://github.com/sovaai/sova-tts
- https://huggingface.co/bene-ges/tts_ru_hifigan_ruslan
- https://github.com/alphacep/vosk-tts
- https://github.com/RHVoice
- https://github.com/snakers4/silero-models#text-to-speech
- https://github.com/Tera2Space/TeraTTS
- https://huggingface.co/omogr/xtts-ru-ipa
- https://www.weights.gg/ru - куча моделей для RVC
- https://2ch-ai.gitgud.site/wiki/speech/ - школота да дваче
- https://lunaiproject.uwu.ai/ - русский DiffSinger
- есть куча телеграм каналов, в основном мутной направленности
- https://github.com/sovaai/sova-tts-tps
- https://github.com/snakers4/silero-models#text-enhancement
- https://github.com/snakers4/russian_stt_text_normalization
- https://www.kaggle.com/competitions/text-normalization-challenge-russian-language/overview - старое соревнование на Kaggle
- https://github.com/ppleskov/Text-Normalization-Challenge-Russian-Language - один из победителей
- https://github.com/shigabeev/russian_tts_normalization
- https://github.com/saarus72/text_normalization/tree/dev - на основе Fred-T5
- https://github.com/Den4ikAI/runorm - числа в текст, обработка английских слов, раскрытие сокращений
- https://github.com/just-ai/multilingual-text-parser
- https://github.com/reynoldsnlp/udar
- https://github.com/einhornus/russian_accentuation
- https://github.com/wilpert/RusPhonetizer
- https://huggingface.co/bene-ges/ru_g2p_ipa_bert_large
- https://github.com/Desklop/StressRNN
- https://github.com/nsu-ai/russian_g2p
- https://github.com/nsu-ai-team/russian_g2p_neuro
- https://github.com/suralmasha/RuTranscript
- https://github.com/MashaPo/russtress
- https://huggingface.co/IlyaGusev/ru-word-stress-transformer
- https://github.com/aishutin/rustress
- https://github.com/Koziev/StressModel
- https://github.com/omogr/omogre
- https://github.com/Den4ikAI/ruaccent - ёфикатор, ударение и разрешение омографов
- https://github.com/reynoldsnlp/udar/blob/main/src/udar/resources/src/Tixonov.txt - Морфемно-орфографический словарь Тихонова
- http://aot.ru - Источник словаря Зализняка в машинном формате
- https://github.com/gramdict/gramdict - современная версия словаря Зализняка
- http://odict.ru/ - другое развитие Зализняка
- http://opencorpora.org/ - размеченный морфологический словарь
- https://ru.wiktionary.org - Wiktionary
- https://kaikki.org/dictionary/Russian/ - дамп wiktionary в удобном формате
- https://github.com/sovaai/sova-tts-tps
- https://github.com/e2yo/eyo-kernel
- https://github.com/kalashnikovisme/karamzin
- https://github.com/Text-extend-tools/python-yoficator
- https://github.com/emacsmirror/yoficator
- https://github.com/unabashed/yoficator
- https://github.com/aniemore/Aniemore
- https://huggingface.co/xbgoose/hubert-large-speech-emotion-recognition-russian-dusha-finetuned
- https://github.com/salute-developers/golos/tree/master/dusha
Сравнение моделей тут.
- Vosk Small https://alphacephei.com/vosk/models/vosk-model-small-ru-0.22.zip
- Vosk Big 0.22 https://alphacephei.com/vosk/models/vosk-model-ru-0.22.zip
- Vosk Big 0.42 https://alphacephei.com/vosk/models/vosk-model-ru-0.42.zip
- Nvidia RNNT Large https://huggingface.co/nvidia/stt_ru_conformer_transducer_large
- Whisper medium https://github.com/openai/whisper
- Whisper Adapted Medium https://huggingface.co/mitchelldehaven/whisper-medium-ru
- Whisper Adapted Large https://huggingface.co/mitchelldehaven/whisper-large-v2-ru
- Wav2VecLM https://huggingface.co/jonatasgrosman/wav2vec2-xls-r-1b-russian
- Wav2VecLM Bond005 https://huggingface.co/bond005/wav2vec2-large-ru-golos (version 03.2023)
- Salute Citrinet https://github.com/salute-developers/golos
- FunASR Russian https://modelscope.cn/models/damo/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/summary
Не тестировались (похуже качеством)
-
https://alphacephei.com/vosk/models/vosk-recasepunc-en-0.22.zip
-
https://github.com/denis-berezutskiy-lad/transcription-bert-ru-punctuator-scripts HuggingFace
-
https://huggingface.co/ai-forever/sage-fredt5-distilled-95m - набор моделей SAGE
