Closed PatWalters closed 1 year ago
What's your purpose of tuning a pre-trained generative model?
I want to fine-tune the pre-trained model to output molecules meeting a specific profile. For instance, outputting molecules that score highly using a target-specific regression model.
I see. DGL-LifeSci does not support that. You may want to check the original codebase of JTVAE or other generative models.
Thanks for your quick response. I'm just curious about the use case for the pre-trained generative models.
No problems! I'll close the issue for now. Feel free to reopen it if you have any further questions.
Is there an example showing how one can use an ML model to tune the output of the pre-trained generative models?