nowordsxiaomu
/

DeepSeek-Flash-Mini

Eval Results (legacy)

Model card Files Files and versions

DeepSeek-Flash-Mini / dataio /__init__.py

nowordsxiaomu's picture

Initial release: DeepSeek-Flash-Mini nano (15M MoE, MLA+MTP)

5e6d9f5 verified about 22 hours ago

History Blame Contribute Delete

200 Bytes

	from .tokenizer import load_tokenizer, train_bpe, ByteTokenizer, BPETokenizer
	from .dataset import BinDataset

	__all__ = ["load_tokenizer", "train_bpe", "ByteTokenizer", "BPETokenizer", "BinDataset"]