You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is an excellent piece of work—very impressive and impactful.
I’m interested in fine-tuning the model using my own dataset (which includes DFT-calculated structures of molecules not present in the original training set)
Would you happen to have any guidance or instructions on how to approach fine-tuning?
Additionally, do you have any intuition or experience regarding how much data is typically required to observe meaningful improvements in performance?
This is an excellent piece of work—very impressive and impactful.
I’m interested in fine-tuning the model using my own dataset (which includes DFT-calculated structures of molecules not present in the original training set)
Would you happen to have any guidance or instructions on how to approach fine-tuning?
Additionally, do you have any intuition or experience regarding how much data is typically required to observe meaningful improvements in performance?