This benchmark compares the efficiency and accuracy of ...
| Program | Language | Framework | Gradient | BITO |
|---|---|---|---|---|
| physher | C | analytic | ||
| phylostan | Stan | Stan | autodiff | |
| phylojax | python | JAX | autodiff | |
| torchtree | python | PyTorch | autodiff | yes |
| treeflow | python | TensorFlow | autodiff | yes |
The gradient of the tree likelihood is optionaly computed by BITO, an efficient C++ library that analytically calculate the gradient using the BEAGLE library. BITO is only available in treeflow and torchtree.
You will need to install nextflow and docker to run this benchmark. Docker is not required but it is highly recommended to use it due to the numerous dependencies.
git clone 4ment/autodiff-experiments.git
git submodule update --init --recursive
nextflow run 4ment/autodiff-experiments -profile docker