Gradient checks for PEtab benchmark models

Gradient checks for the collection of PEtab benchmark models. The gradients checked are performed using AMICI for simulation, via interfaces in AMICI and pyPESTO.

Usage

1. Link to the benchmark models. Either:
   • clone the models inside this repository with

   git clone https://github.com/Benchmarking-Initiative/Benchmark-Models-PEtab.git

   • or, change paths to match a pre-existing copy of the models
     • _helpers.py: benchmark_path
     • run.sh: benchmark_dir
2. Setup a Python environment with:
   • AMICI
   • pyPESTO
   • libpetab-python
3. Submit the SLURM jobs with bash run.sh
4. Plot results with python3 plot.py

TODO

• tee output to log file named by tool/model/vector
• parallelize across test vectors
• use job array
• add number of failing parameters to each cell in the plot heatmap