14 10 2024 chemoinformatics Tweet
As it could not be installed with the original requirements, I modified it and uploaded it to the Mishima-syk repository.
$ git clone https://github.com/Mishima-syk/REINVENT4.git $ cd REINVENT4 $ conda create --name reinvent4 python=3.11 $ conda activate reinvent4 $ pip install -r requirements-macOS.lock $ pip install --no-deps .
Now, you can set mps (Metal Performance Shaders) in the device parameter
# mps_sampling.toml # REINVENT4 TOML input example for sampling run_type = "sampling" device = "mps" # M3 GPU json_out_config = "_sampling.json" # write this TOML to JSON [parameters] ## Reinvent: de novo sampling model_file = "priors/reinvent.prior" output_file = 'sampling.csv' # sampled SMILES and NLL in CSV format num_smiles = 157 # number of SMILES to be sampled, 1 per input SMILES unique_molecules = true # if true remove all duplicatesd canonicalize smiles randomize_smiles = true # if true shuffle atoms in SMILES randomly
Running the reinvent4 program
$ reinvent -l sampling.log mps_sampling.toml