Here is a quick summary of the merged pull requests for DeepChem during the second week of December:
1. Missing Modules fix: This PR fixes an issue related to missing modules.
2. read the docs fix: This PR addresses the need for passing python.version in the build as readthedocs no longer supports it.
3. OrbParam Shift to deep_dft: This PR fixes an issue related to OrbParam by shifting it to deep_dft.
4. fix read the docs build: This PR adds a required build option to create the documentation in readthedocs.
5. error fix in safeops: This PR fixes issue(#3717) related to an error when importing deepchem.
**6. add vina to python 3.10:**This PR fixes a Vina issue that was preventing tests from running in Python 3.10.
7. Safe Operations: This PR resolves safeops issue and enhances utilities for safe tensor operations for avoiding NaNs and Infs in the output.
8. Fixed missing comma that broke notebook and switched glitter to discord: This PR Addresses a missing comma issue in tutorial notebooks and switches glitter to discord.
9. MulLinearOperator LinearOperator Helper Class: Fixes issue by introducing MulLinearOperator as a helper class for linear operations.
10. Update models init.py: This PR moves MEGNetModel to the torch-geometric dependency section.
11. Added weight decay argument to adam optimizer: This PR introduces weight decay argument to the Adam optimizer.
12. Fix Atomic Convolution Model (ACNN) tests in Pytorch and Tensorflow: Resolves a ValueError in ACNN tensorflow and pytorch0 tests and updates featurization with dtype=object.
13. BaseHamilton Main Class: This PR fixes issue in BaseHamilton main class.
14. remove mypy pin: This PR fixes issue #3674 to remove the mypy pin and plans to update to version 1.7.1 once released.
15. Fix Transformers Pin and Docs: This PR fixes the transformers pin, merges updates into torch-lightning-2.1.2, and resolves a missing GradientPenaltyLayer in the documentation.