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
When the seed molecule contains biaryl or fused rings, atom fitting algorithm generates incorrect number of atoms.
For example, when I tried with 5qax, all generated rings were 5-membered rings.
I guessed that the grid density of these compounds are well overlapped and make it difficult to generate correct number of atoms.
To overcome this problem, I've changed radious_scale of gmaker in train.py and generate.py from 1.0 to 0.8. But it did't work well, because the grid value were very small to put atoms.
Do you have any ideas for this problems?
Another test set are 5d7x and 2x7t.
The text was updated successfully, but these errors were encountered:
When the seed molecule contains biaryl or fused rings, atom fitting algorithm generates incorrect number of atoms.
For example, when I tried with 5qax, all generated rings were 5-membered rings.
I guessed that the grid density of these compounds are well overlapped and make it difficult to generate correct number of atoms.
To overcome this problem, I've changed radious_scale of gmaker in train.py and generate.py from 1.0 to 0.8. But it did't work well, because the grid value were very small to put atoms.
Do you have any ideas for this problems?
Another test set are 5d7x and 2x7t.
The text was updated successfully, but these errors were encountered: