Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use Chemprop version 2 #14

Open
wants to merge 10 commits into
base: main
Choose a base branch
from

Conversation

davide-grheco
Copy link

This update transitions the codebase from Chemprop v1 to Chemprop v2 to address compatibility issues and enhance functionality, as discussed in Issue #11. Given that Chemprop v1 is no longer supported, this upgrade ensures access to ongoing updates and support from the Chemprop community.

Key improvements with Chemprop v2 include:

  • Direct RDKit Mol Object Support: Chemprop v2 introduces the MoleculeDataPoint class, allowing direct initialization from RDKit Mol objects and reducing the need for repetitive SMILES-to-Mol conversions, which streamlines predictions and minimizes computational demands.
  • Enhanced Consistency: By directly supporting Mol objects, this version reduces potential mismatches from converting between SMILES and Mol objects, adding reliability for workflows where Mol objects are already used.

Additionally, the code was refactored into smaller functions to improve readability and modularity. However, if preferred, these changes can be reverted to the previous structure.

Finally, it will be necessary to verify that models with Chemprop v2 perform comparably with those from v1 to ensure accuracy and consistency. Benchmark testing will be conducted to validate model outputs and adjust if needed to align with v1 performance levels.

@swansonk14
Copy link
Owner

Hi @davide-grheco,

Thank you so much for creating this PR and doing all the work to transition to Chemprop v2! I'll aim to work on checking the code and training the Chemprop v2 models in December to ensure that they maintain the same level of accuracy as the v1 models.

Best,
Kyle

@davide-grheco
Copy link
Author

Hello @swansonk14,
thank you for taking a look and performing the validation. I'll be quite busy in December, but if you have any comment or you feel something should be changed feel free to contact me here or at my email.

Greetings
Davide

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants