Til Web, Crawls and Archive Web text, crawls and archived sources for pretraining corpora. Веб-тексты, обходы сайтов и архивные источники для корпусов предобучения. Collection by TilQazyna Aug 9 - TilQazyna/Til-Web-Raw-KK-v1 Preview • Updated Aug 8 • 141 TilQazyna/Til-Zharar Viewer • Updated Aug 8 • 10.1k • 117 TilQazyna/datalake Updated Aug 8 • 731 TilQazyna/datalake-part2 Preview • Updated Aug 8 • 129
Til Instruct — Task Datasets Instruction data for QA, summarization, classification and generation. Данные инструкций для вопросов, суммаризации, классификации и генерации. Collection by TilQazyna Aug 8 - TilQazyna/til-kk-qa-v1 Viewer • Updated Aug 8 • 32.9k • 239 TilQazyna/til-kk-summ-v1 Viewer • Updated Aug 8 • 11.8k • 243 TilQazyna/til-kk-classify-v1 Viewer • Updated Aug 8 • 11.9k • 243 TilQazyna/til-kk-keywords-v1 Viewer • Updated Aug 8 • 12.2k • 239
Til 256k Research Ladder Linked pretraining, SFT and GEC artifacts for the exp080–exp085 sequence. Связанные артефакты предобучения, SFT и GEC для цепочки exp080–exp085. Collection by TilQazyna Aug 8 - TilQazyna/Til-kk-0.5B-256k-exp080 Text Generation • Updated Aug 8 • 2 TilQazyna/Til-kk-0.5B-256k-sft-exp081 Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-exp082 Text Generation • Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-rlaif-exp083 Updated Aug 8
Til — Multilingual Models Kazakh-first multilingual models for text generation and correction. Многоязычные модели с приоритетом казахского для генерации и коррекции. Collection by TilQazyna Aug 8 - TilQazyna/Til-mini-1B Text Generation • 1.0B • Updated Aug 8 • 1 TilQazyna/Til-mini-1B-GEC Text Generation • 1.0B • Updated Aug 8 TilQazyna/Til-2B Text Generation • 2B • Updated Aug 8 TilQazyna/Til-2B-instruct Text Generation • 2B • Updated Aug 8
Til Flagship Datasets Core datasets for pretraining, instruction tuning, speech and language tasks. Основные датасеты для предобучения, инструкций, речи и языковых задач. Collection by TilQazyna Aug 9 - TilQazyna/Til-Corpus Updated Aug 9 • 498 • 1 TilQazyna/Til-Instruct Viewer • Updated Aug 8 • 6.26M • 245 TilQazyna/Til-Parallel Updated Aug 8 • 120 TilQazyna/Til-Audio Viewer • Updated Aug 8 • 380k • 234
Til Books — Sources and Scans Extracted books, scans and source subsets for corpus builders. Извлечённые книги, сканы и наборы по источникам для сборки корпусов. Collection by TilQazyna Aug 9 -
Tokenizers Tokenizer artifacts and vocabulary data for Kazakh model training. Tokenizer и словарные данные для обучения моделей казахского языка. Collection by TilQazyna Aug 8 - TilQazyna/Til-Tokenizer-128k Updated Aug 8 TilQazyna/Til-vocab-morphbpe-96k-kkrumix-v1 Updated Aug 8 • 127
Kazakh GEC — All Models Kazakh correction models and training data for GEC comparison. Модели коррекции казахского текста и данные для сравнения GEC. Collection by TilQazyna Aug 8 - TilQazyna/Til-GEC-Gemma4-exp087 Text Generation • Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-exp085 Text Generation • 0.8B • Updated Aug 8 TilQazyna/Til-2B-GEC Text Generation • 2B • Updated Aug 8 • 20 TilQazyna/Til-mini-1B-GEC Text Generation • 1.0B • Updated Aug 8
Til Core — Kazakh-only Models Kazakh-only base, instruct and GEC models for researchers and developers. Базовые, instruct- и GEC-модели только для казахского языка. Collection by TilQazyna Aug 8 - TilQazyna/Til-Core-0.5B Text Generation • 0.5B • Updated Aug 8 • 4 TilQazyna/Til-Core-0.5B-Instruct Text Generation • 0.5B • Updated Aug 8 • 1 TilQazyna/Til-Core-0.5B-QazGramma-GEC Text Generation • 0.5B • Updated Aug 8 TilQazyna/Til-Core-1B Text Generation • 2B • Updated Aug 8
Useless Voice Collection by lepwnr Aug 8 - Running on Zero Agents Featured 2.94k F5-TTS 🗣 2.94k F5-TTS & E2-TTS: Zero-Shot Voice Cloning (Unofficial Demo)
Running on Zero Agents Featured 2.94k F5-TTS 🗣 2.94k F5-TTS & E2-TTS: Zero-Shot Voice Cloning (Unofficial Demo)
Til Web, Crawls and Archive Web text, crawls and archived sources for pretraining corpora. Веб-тексты, обходы сайтов и архивные источники для корпусов предобучения. Collection by TilQazyna Aug 9 - TilQazyna/Til-Web-Raw-KK-v1 Preview • Updated Aug 8 • 141 TilQazyna/Til-Zharar Viewer • Updated Aug 8 • 10.1k • 117 TilQazyna/datalake Updated Aug 8 • 731 TilQazyna/datalake-part2 Preview • Updated Aug 8 • 129
Til Books — Sources and Scans Extracted books, scans and source subsets for corpus builders. Извлечённые книги, сканы и наборы по источникам для сборки корпусов. Collection by TilQazyna Aug 9 -
Til Instruct — Task Datasets Instruction data for QA, summarization, classification and generation. Данные инструкций для вопросов, суммаризации, классификации и генерации. Collection by TilQazyna Aug 8 - TilQazyna/til-kk-qa-v1 Viewer • Updated Aug 8 • 32.9k • 239 TilQazyna/til-kk-summ-v1 Viewer • Updated Aug 8 • 11.8k • 243 TilQazyna/til-kk-classify-v1 Viewer • Updated Aug 8 • 11.9k • 243 TilQazyna/til-kk-keywords-v1 Viewer • Updated Aug 8 • 12.2k • 239
Tokenizers Tokenizer artifacts and vocabulary data for Kazakh model training. Tokenizer и словарные данные для обучения моделей казахского языка. Collection by TilQazyna Aug 8 - TilQazyna/Til-Tokenizer-128k Updated Aug 8 TilQazyna/Til-vocab-morphbpe-96k-kkrumix-v1 Updated Aug 8 • 127
Til 256k Research Ladder Linked pretraining, SFT and GEC artifacts for the exp080–exp085 sequence. Связанные артефакты предобучения, SFT и GEC для цепочки exp080–exp085. Collection by TilQazyna Aug 8 - TilQazyna/Til-kk-0.5B-256k-exp080 Text Generation • Updated Aug 8 • 2 TilQazyna/Til-kk-0.5B-256k-sft-exp081 Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-exp082 Text Generation • Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-rlaif-exp083 Updated Aug 8
Kazakh GEC — All Models Kazakh correction models and training data for GEC comparison. Модели коррекции казахского текста и данные для сравнения GEC. Collection by TilQazyna Aug 8 - TilQazyna/Til-GEC-Gemma4-exp087 Text Generation • Updated Aug 8 TilQazyna/Til-kk-0.5B-256k-gec-exp085 Text Generation • 0.8B • Updated Aug 8 TilQazyna/Til-2B-GEC Text Generation • 2B • Updated Aug 8 • 20 TilQazyna/Til-mini-1B-GEC Text Generation • 1.0B • Updated Aug 8
Til — Multilingual Models Kazakh-first multilingual models for text generation and correction. Многоязычные модели с приоритетом казахского для генерации и коррекции. Collection by TilQazyna Aug 8 - TilQazyna/Til-mini-1B Text Generation • 1.0B • Updated Aug 8 • 1 TilQazyna/Til-mini-1B-GEC Text Generation • 1.0B • Updated Aug 8 TilQazyna/Til-2B Text Generation • 2B • Updated Aug 8 TilQazyna/Til-2B-instruct Text Generation • 2B • Updated Aug 8
Til Core — Kazakh-only Models Kazakh-only base, instruct and GEC models for researchers and developers. Базовые, instruct- и GEC-модели только для казахского языка. Collection by TilQazyna Aug 8 - TilQazyna/Til-Core-0.5B Text Generation • 0.5B • Updated Aug 8 • 4 TilQazyna/Til-Core-0.5B-Instruct Text Generation • 0.5B • Updated Aug 8 • 1 TilQazyna/Til-Core-0.5B-QazGramma-GEC Text Generation • 0.5B • Updated Aug 8 TilQazyna/Til-Core-1B Text Generation • 2B • Updated Aug 8
Til Flagship Datasets Core datasets for pretraining, instruction tuning, speech and language tasks. Основные датасеты для предобучения, инструкций, речи и языковых задач. Collection by TilQazyna Aug 9 - TilQazyna/Til-Corpus Updated Aug 9 • 498 • 1 TilQazyna/Til-Instruct Viewer • Updated Aug 8 • 6.26M • 245 TilQazyna/Til-Parallel Updated Aug 8 • 120 TilQazyna/Til-Audio Viewer • Updated Aug 8 • 380k • 234
Useless Voice Collection by lepwnr Aug 8 - Running on Zero Agents Featured 2.94k F5-TTS 🗣 2.94k F5-TTS & E2-TTS: Zero-Shot Voice Cloning (Unofficial Demo)
Running on Zero Agents Featured 2.94k F5-TTS 🗣 2.94k F5-TTS & E2-TTS: Zero-Shot Voice Cloning (Unofficial Demo)