carate 0.3.9

Providing reproducible modeling

CARATE

• 1. Why
• 2. What
• 3. Scope
• 4. Installation on CPU
  • 4.1. Build manually
  • 4.2. Installation from repo
• 5. Usage
  • 5.1. Start a run
  • 5.2. Training results
• 6. Build on the project
• 7. Support the development
• 8. Cite

1. Why

Molecular representation is wrecked. Seriously! We chemists talked for decades with an ancient language about something we can't comprehend with that language. We have to stop it, now!

2. What

The success of transformer models is evident. Applied to molecules we need a graph-based transformer. Such models can then learn hidden representations of a molecule better suited to describe a molecule.

For a chemist it is quite intuitive but seldomly modelled as such: A molecule exhibits properties through its combined electronic and structural features

• Evidence of this perspective was given in chembee.

• Mathematical equivalence of the variational principle and neural networks was given in the thesis Markov-chain modelling of dynmaic interation patterns in supramolecular complexes.

• The failure of the BOA is described in the case of diatomic tranistion metal fluorides is described in the preprint: Can Fluorine form triple bonds?

• Evidence of quantum-mechanical simulations via molecular dynamics is given in a seminal work Direct Simulation of Bose-Einstein-Condensates using molecular dynmaics and the Lennard-Jones potential

3. Scope

The aim is to implement the algorithm in a reusable way, e.g. for the chembee pattern. Actually, the chembee pattern is mimicked in this project to provide a stand alone tool. The overall structure of the program is reusable for other deep-learning projects and will be transferred to an own project that should work similar to opinionated frameworks.

4. Installation on CPU

Prepare system

sudo apt-get install python3-dev libffi-dev

4.1. Build manually

pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu 
pip install torch-scatter torch-sparse torch-geometric rdkit-pypi networkx[default] matplotlib
pip install torch-cluster 
pip install torch-spline-conv 

4.2. Installation from repo

Inside the directory of your git-clone:

pip install -e .

5. Usage

The program is meant to be run as a simple CLI. You can specify the configuration either via a JSON and use the program as a microservice, or you may run it locally from the command line. It is up to you.

Finally, with the new pyproject.toml it is possible to

pip install carate

The installation will install torch with CUDA, so the decision of the library what hardware to use goes JIT (just-in-time). At the moment only CPU/GPU is implemented and FPGA/TPU and others are ignored. Further development of the package will then focus on avoiding special library APIs but make the pattern adaptable to an arbitrary algorithmic/numerical backend.

carate -c path_to_config_file.py

5.1. Start a run

To start a run you need to define the configuration. You can do so by defining a .json or a config.py file

Examples for config.py files are given in config_files

Or you can check the the tutorial.ipynb in notebooks how to use the package with a .json file

5.2. Training results

Most of the training results are saved in a accumulative json on the disk. The reason is to have enough redundancy in case of data failure.

Previous experiments suggest to harden the machine for training to avoid unwanted side-effects as shutdowns, data loss, or data diffusion. You may still send intermediate results through the network, but store the large chunks on the hardened device.

Therefore, any ETL or data processing might not be affected by any interruption on the training machine.

6. Build on the project

Building on the code is not recommended as the project will be continued in another library (building with that would make most sense).

However, you may still use the models as they are by the means of the library production ready.

In case you can't wait for the picky scientist in me, you can still build on my intermediate results. You can find them in the following locations

• Google Drive

We have to admit it though: There was a security incident on 31st of March 2023, so the results from Alchemy and ZINC are still waiting. I logged all experiments I did and uploaded the log, such that any picky reviewer can check where there happened weird stuff, I made a mistake, or there was an incident and requests more reproductions. That should solve the issue for now!

7. Support the development

If you are happy about substantial progress in chemistry and life sciences that is not commercial first but citizen first, well then just

8. Cite

There is a preprint available on bioRxiv. Read the preprint