Are these the quantized versions of the Instruct models or the base models ?
#1
by
manjunaths
- opened
Basically the title, are these the quantized versions of the instruct models or the base models ?
I ask because none of the models have the instruct in the file names.
Thanks.
Hey @manjunaths , these are quantized versions for base model. The tokenizer for our instruct model has issues and we're trying to figure out: https://github.com/ggerganov/llama.cpp/pull/10535