how to use GPU

[wenchangzhou-qtx]

Hey @michael1788,

Thanks for such a nice mode for generating molecules! I tried to run the experiment but found it was very slow, the training step gives me some error (see below) which seems suggest I don't have GPU, but I do have one, so question is how to set it up so that I can make use of my GPU to acerbate the generating process, otherwise I have to spend a few days to run one experiment. Can you suggest something? Thanks!

I clone and install hybridCLMs according to your instruction, one thing I found is the tensorflow does not have gpu support, is that the case? Or maybe tensorflow by default supports GPU?

tensorboard               2.2.2                    pypi_0    pypi
tensorboard-plugin-wit    1.8.1                    pypi_0    pypi
tensorflow                2.2.0                    pypi_0    pypi
tensorflow-estimator      2.2.0                    pypi_0    pypi

Output message when I run the training step

$ sh run_training.sh configfiles/ft_clm_generation/A01_clm_ft.ini
Using TensorFlow backend.

START TRAINING

Batch_size used: 4
Data path : ../data/fine_tuning_generation/1_90_x0/
2023-10-24 10:30:19.381754: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2023-10-24 10:30:19.447307: E tensorflow/stream_executor/cuda/cuda_driver.cc:313] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected
2023-10-24 10:30:19.447328: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (login1): /proc/driver/nvidia/version does not exist
2023-10-24 10:30:19.447478: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 AVX512F FMA
2023-10-24 10:30:19.465903: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 2300000000 Hz
2023-10-24 10:30:19.468411: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x1552b0000b60 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2023-10-24 10:30:19.468428: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
Weights loaded: ../pretrained_models/CLM.h5
Model: "sequential_1"
.....