Installing Dependencies

Before training the DLRM model, install the third-party libraries h5py, TensorFlow 2.13.0, and tqdm for data preprocessing, model training, and inference.

1. Enter the conda base environment.

   conda activate

2. Create a conda virtual environment.

   conda create -n  dlrm  python=3.9

   

   If the preceding information is displayed, the conda virtual environment is successfully created.

   

   • -n: specifies the environment name.

   • python=3.9: specifies the Python version.
3. Switch to the newly created conda virtual environment dlrm.

   conda activate dlrm

4. Verify the Python version of the new virtual environment.

   python -V 

   

   If Python 3.9.19 is displayed, Python is correctly installed.

5. Install the third-party library h5py.

   conda install h5py=3.11.0

   

   If the preceding information is displayed, the installation is successful.

6. Install TensorFlow 2.13.0 and tqdm.

   pip install tensorflow==2.13.0 tqdm==4.66.2

   

   If the preceding information is displayed, the installation is successful.