Phi3 or Mistral?
#3
by
csabakecskemeti
- opened
Why the architecture states
"MistralForCausalLM"?
Is this made from microsoft/Phi-3-medium-4k-instruct ?
Why the architecture states
"MistralForCausalLM"?
Is this made from microsoft/Phi-3-medium-4k-instruct ?
Hi there so this model is microsoft/Phi-3-medium-4k-instruct however we just converted it to Mistral architecture so it is more accurate, efficient and easier to use. regarding training, serving etc.
Ahh ok, nice!
csabakecskemeti
changed discussion status to
closed