Fine tuning in my own domain

#1
by wilfoderek - opened

Guys really amazing work! brave!!!
Now i have a questiosn for all of you. I want to fine tuning this model for my own domain, so could you explain how to achieve this ?
Thanks in advance

embaas org

Hey,

all credits go to intfloat and Microsoft. @intfloat mentioned his training here: https://github.com/microsoft/unilm/tree/master/simlm.

But the model you work also with sentence-transformers training functions.

Another possibility would be just to fine tune the last layer, which seems to be a easy and powerful solution. Check out this here.

Man, thank you so much! i will check out "Another possibility would be just to fine tune the last layer". i

Banso changed discussion status to closed

Sign up or log in to comment