Why "Mistral Inference" and why your own tokenizer formats?
#34
by
winddude
- opened
Curious, what are the advantages of "mistral inference" and using your tokenizer formats, eg, "tokenizer.model.v3" over huggingface transformers?
Curious, what are the advantages of "mistral inference" and using your tokenizer formats, eg, "tokenizer.model.v3" over huggingface transformers?