Aron Jansen

**Describe the bug** Running `pip install deeprank` I have an issue with the freesasa dependency, see full error below. **Environment:** - OS system: M1 Mac - Version: Monterey 12.4 **Actual...

To test the explanations of SMILES based models, we'd like a model that predicts some simple molecular property for which an explanation at the level of atoms makes sense and...

Ignore 10% worst replicas in hyper loss

3

Basic implementation, not tested yet and the percentage is not configurable from the runcard.

NaNs in hyperopt

8

I created a small script that gets the parameters from a trial and creates a runcard from that [here](https://github.com/LHC-NLeSC/nnpdf-scripts/blob/main/hyperopt-final/create_trial_runcard.py) (where I means GPT4). ## Side issue, bug? Regardless of using...

Avoid idle gpu

1

# The idea We observed large gaps between training steps in the tensorboard profile when running on the GPU. After a lot of fiddling were found to be (at least...

Finalizing eScience contributions

1

# Finalizing eScience contributions Since our time is running out, we thought it would be useful to have an overview of our remaining PRs, separated into essentials and optionals. I...

Avoiding duplicated computations by having a single observable model

34

# Goal The goal of this PR is to speed up the code by a factor 2 by a refactoring that avoids redoing the same computations. Currently there are separate...

Realising a factor 20-30 speedup on GPU

3

Last week @goord and I started looking at tensorboard profiles of the code running on a GPU. We found and resolved several bottlenecks in the performance, resulting in a total...