Where is the tokenizer?

#2
by EarthWorm001 - opened

I'm trying to run the Mamba-2 hybrid using megatron, However, the example scripts require a tokenizer prepared. Where to download the tokenizer?

Thanks!

NVIDIA org

The tokenizer is uploaded in each repository under the filename mt_nlg_plus_multilingual_ja_zh_the_stack_frac_015_256k.model

rwaleffe changed discussion status to closed

Sign up or log in to comment