Is this only for loading the nemotron tokenizer?
#1
by
ZQ-Dev
- opened
This repo seems to make the tokenizer load-able with HF transformers, but what about the model itself. Is this tokenizer fix all you need to load the model with AutoModelForCausalLM? If not, what are the next steps to work towards that?
Would be great to one day load nemotron-4 with TGI.
how can i use the nemotron for the data generation ??