INVALID LANGUAGE PAIR SPECIFIED. EXAMPLE: LANGPAIR=EN|IT USING 2 LETTER ISO OR RFC3066 LIKE ZH-CN. ALMOST ALL LANGUAGES SUPPORTED BUT SOME MAY HAVE NO CONTENT Any build from master after that PR works. Wrap DNA sequences in tags so the tokenizer switches into 6-mer mode: — DNA tag is mandatory. Raw DNA without the tag will be BPE-tokenized as English text, producing meaningless results. — 6-mer constraint. DNA sequences must contain only [ATCG] characters. Any other letters trigger an token. — fns revision is not supported. This GGUF is converted from the main revision only. The fns branch uses a factorized nucleotide supervision head that requires the Python transformers path. — Carbon-500M** (GGUF here) can be used as a draft model for speculative decoding to speed up Carbon-3B inference at no quality loss. These quants were produced from the original bf16…
Модальности:
Генерация текста
Области применения:
Биология
Задача: Генерация текста
Автор: pankajpandey-dev
Теги: gguf, biology, carbon, dna, genomics, hybrid-tokenizer, llama.cpp, ollama
Лайков: 3 | Загрузок: 258
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
