MOFA2
MOFA2 copied to clipboard
bioFAM
How to specify GPU device in MOFA2
Hi,
I run MOFA2 in R with reticulate. I have 2 GPUs in my machine, do you know how to specify the GPU device to run MOFA2?
Usually, in conda environment, I can set up the GPU device by export CUDA_VISIBLE_DEVICES=1,0 or CUDA_VISIBLE_DEVICES=1 ./cuda_executable
But I don't know how to do it in R.
Hey @yulijia,
I think you can still set environmental variables in R.
Not sure if that is going to work
Would a train_opts$gpu_device be something that you consider a good solution for your use case?
Hi @gtca , Thank you for your quick response! Unfortunately, set environmental variables in R doesn't work in Rstudio server.
I think train_opts$gpu_device may the best choice for specific GPU devices in the R package.
If it is possible, please also consider something like model = nn.DataParallel(model).cuda() to implement data parallelism on all GPU devices when the memory of single GPU is not enough to train the model.
Hi @gtca
I recently updated the mofapy2 package (0.6.4 -> 0.7.0) on my server and run the package using mofa r package with GPU.
However, it is never run on GPU anymore.
I have installed Cupy, CUDA Toolkit, dwf-python, and mofapy2. However, it seems that regardless of the specified gpu_device number, the code is defaulting to the CPU version. But Previous version, I ran it on GPU successfully.
Do you know how to solve the problem?
Here is my code to run mofa.
## cbmc is a SeuratData, https://github.com/satijalab/seurat-data
MOFAobject <- create_mofa(cbmc)
data_opts <- get_default_data_options(MOFAobject)
model_opts <- get_default_model_options(MOFAobject)
train_opts <- get_default_training_options(MOFAobject)
train_opts$gpu_mode <- TRUE
train_opts$gpu_device=1
MOFAobject <- prepare_mofa(
object = MOFAobject,
data_options = data_opts,
model_options = model_opts,
training_options = train_opts
)
keyword="cbmc"
outfile = paste0("./",keyword,"_model.hdf5")
use_condaenv(condaenv = "mofa2",conda="/home/ylj/miniconda3/bin/conda",required=T)
use_python("/home/ylj/miniconda3/envs/mofa2/bin/python3.8",required = T)
py_module_available("mofapy2") ##return TRUE
.mofapy2_version <- "0.7.0"
MOFA2:::.run_mofa_reticulate(MOFAobject, outfile, save_data=TRUE)## throw error at this line
MOFAobject.trained <- load_model(outfile)
factors <- get_factors(MOFAobject.trained, factors = "all")
Here are the packages that in my conda env
_libgcc_mutex 0.1 conda_forge conda-forge
_openmp_mutex 4.5 2_gnu conda-forge
anndata 0.8.0 pypi_0 pypi
argparse 1.4.0 pypi_0 pypi
brotli-python 1.1.0 py38h17151c0_1 conda-forge
bzip2 1.0.8 hd590300_5 conda-forge
ca-certificates 2023.11.17 hbcca054_0 conda-forge
certifi 2023.11.17 pyhd8ed1ab_0 conda-forge
charset-normalizer 3.3.2 pyhd8ed1ab_0 conda-forge
cudatoolkit 11.8.0 h4ba93d1_12 conda-forge
cupy-cuda110 12.2.0 pypi_0 pypi
dtw-python 1.3.0 py38h01eb140_2 conda-forge
fastrlock 0.8.2 pypi_0 pypi
h5py 3.10.0 pypi_0 pypi
idna 3.5 pyhd8ed1ab_0 conda-forge
joblib 1.3.2 pypi_0 pypi
ld_impl_linux-64 2.40 h41732ed_0 conda-forge
libblas 3.9.0 20_linux64_openblas conda-forge
libcblas 3.9.0 20_linux64_openblas conda-forge
libffi 3.4.2 h7f98852_5 conda-forge
libgcc-ng 13.2.0 h807b86a_3 conda-forge
libgfortran-ng 13.2.0 h69a702a_3 conda-forge
libgfortran5 13.2.0 ha4646dd_3 conda-forge
libgomp 13.2.0 h807b86a_3 conda-forge
liblapack 3.9.0 20_linux64_openblas conda-forge
libnsl 2.0.1 hd590300_0 conda-forge
libopenblas 0.3.25 pthreads_h413a1c8_0 conda-forge
libsqlite 3.44.2 h2797004_0 conda-forge
libstdcxx-ng 13.2.0 h7e041cc_3 conda-forge
libuuid 2.38.1 h0b41bf4_0 conda-forge
libzlib 1.2.13 hd590300_5 conda-forge
mofapy2 0.7.0 pypi_0 pypi
natsort 8.4.0 pypi_0 pypi
ncurses 6.4 h59595ed_2 conda-forge
numpy 1.24.4 py38h59b608b_0 conda-forge
openssl 3.2.0 hd590300_0 conda-forge
packaging 23.2 pyhd8ed1ab_0 conda-forge
pandas 1.5.3 pypi_0 pypi
pip 23.3.1 pyhd8ed1ab_0 conda-forge
platformdirs 4.0.0 pyhd8ed1ab_0 conda-forge
pooch 1.8.0 pyhd8ed1ab_0 conda-forge
pysocks 1.7.1 pyha2e5f31_6 conda-forge
python 3.8.18 hd12c33a_0_cpython conda-forge
python-dateutil 2.8.2 pypi_0 pypi
python_abi 3.8 4_cp38 conda-forge
pytz 2023.3.post1 pypi_0 pypi
readline 8.2 h8228510_1 conda-forge
requests 2.31.0 pyhd8ed1ab_0 conda-forge
scikit-learn 1.3.2 pypi_0 pypi
scipy 1.10.1 py38h59b608b_3 conda-forge
setuptools 68.2.2 pyhd8ed1ab_0 conda-forge
six 1.16.0 pypi_0 pypi
threadpoolctl 3.2.0 pypi_0 pypi
tk 8.6.13 noxft_h4845f30_101 conda-forge
typing_extensions 4.8.0 pyha770c72_0 conda-forge
urllib3 2.1.0 pyhd8ed1ab_0 conda-forge
wheel 0.41.3 pyhd8ed1ab_0 conda-forge
xz 5.2.6 h166bdaf_0 conda-forge
Hi @gtca ,
Problem solved, i forget to install cupy...
so the correct packages in env should includes:
_libgcc_mutex 0.1 conda_forge conda-forge
_openmp_mutex 4.5 2_gnu conda-forge
anndata 0.9.2 pypi_0 pypi
bzip2 1.0.8 hd590300_5 conda-forge
ca-certificates 2023.11.17 hbcca054_0 conda-forge
cuda-version 11.8 h70ddcb2_2 conda-forge
cudatoolkit 11.8.0 h4ba93d1_12 conda-forge
cupy 12.2.0 py38h7b7cd4b_4 conda-forge
cupy-cuda11x 12.2.0 pypi_0 pypi
fastrlock 0.8.2 py38h17151c0_1 conda-forge
h5py 3.10.0 pypi_0 pypi
joblib 1.3.2 pypi_0 pypi
ld_impl_linux-64 2.40 h41732ed_0 conda-forge
libblas 3.9.0 20_linux64_openblas conda-forge
libcblas 3.9.0 20_linux64_openblas conda-forge
libffi 3.4.2 h7f98852_5 conda-forge
libgcc-ng 13.2.0 h807b86a_3 conda-forge
libgfortran-ng 13.2.0 h69a702a_3 conda-forge
libgfortran5 13.2.0 ha4646dd_3 conda-forge
libgomp 13.2.0 h807b86a_3 conda-forge
liblapack 3.9.0 20_linux64_openblas conda-forge
libnsl 2.0.1 hd590300_0 conda-forge
libopenblas 0.3.25 pthreads_h413a1c8_0 conda-forge
libsqlite 3.44.2 h2797004_0 conda-forge
libstdcxx-ng 13.2.0 h7e041cc_3 conda-forge
libuuid 2.38.1 h0b41bf4_0 conda-forge
libzlib 1.2.13 hd590300_5 conda-forge
mofapy2 0.7.1 pypi_0 pypi
natsort 8.4.0 pypi_0 pypi
ncurses 6.4 h59595ed_2 conda-forge
numpy 1.24.4 py38h59b608b_0 conda-forge
openssl 3.2.0 hd590300_0 conda-forge
packaging 23.2 pypi_0 pypi
pandas 2.0.3 pypi_0 pypi
pip 23.3.1 pyhd8ed1ab_0 conda-forge
python 3.8.18 hd12c33a_0_cpython conda-forge
python-dateutil 2.8.2 pypi_0 pypi
python_abi 3.8 4_cp38 conda-forge
pytz 2023.3.post1 pypi_0 pypi
readline 8.2 h8228510_1 conda-forge
scikit-learn 1.3.2 pypi_0 pypi
scipy 1.10.1 pypi_0 pypi
setuptools 68.2.2 pyhd8ed1ab_0 conda-forge
six 1.16.0 pypi_0 pypi
threadpoolctl 3.2.0 pypi_0 pypi
tk 8.6.13 noxft_h4845f30_101 conda-forge
tzdata 2023.3 pypi_0 pypi
wheel 0.41.3 pyhd8ed1ab_0 conda-forge
xz 5.2.6 h166bdaf_0 conda-forge