> This page is for version 10. > For other versions, use one of these documentation indexes: > - 11 (default): https://docs.nvidia.com/base-command-manager/product-manuals/11/llms.txt > - 10: https://docs.nvidia.com/base-command-manager/product-manuals/10/llms.txt > For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.nvidia.com/base-command-manager/product-manuals/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/base-command-manager/product-manuals/_mcp/server. # Running TensorFlow This chapter goes through some example runs with TensorFlow. Some output messages have been removed or simplified in the runs for readability. The sample runs assume that the GPU drivers are in place ([section 9.1 of the *Installation Manual*](/installation/nvidiagpus#91-installing-cuda)). The runs also assume that TensorFlow package and the TensorFlow extras package have both been installed from the NVIDIA Base Command Manager repository with a package manager. TensorFlow can be installed, for example with `yum install cm-tensorflow2-py39-cuda11.2-gcc9 cm-tensorflow2-extra-py39-cuda11.2-gcc9` In addition to requiring the extras package, TensorFlow requires an OpenMPI implementation to work. [Chapter 3](/user/mpi) of the *User Manual* describes the different OpenMPI packages that the BCM repositories provide. The different OpenMPI packages allow the user to choose which one to use. For example, depending on which interconnect is being used, or depending on if CUDA support is required. In this chapter, the `cm-openmpi4-cuda11.2-ofed51-gcc9` package is used. It can be installed with, for example: `yum install cm-openmpi4-cuda11.2-ofed51-gcc9` More information on the examples can be found at [https://github.com/tensorflow/examples](https://github.com/tensorflow/examples). ## 2.1 Hello World A “`Hello World`” interactive example, that just shows that the software is in place for TensorFlow, can be run as follows: **Example** ```bash [root@basecm10 ~]# module load shared [root@basecm10 ~]# module load tensorflow2-py39-cuda11.2-gcc9 Loading tensorflow2-py39-cuda11.2-gcc9/2.7.0 Loading requirement: openblas/dynamic/0.3.18 hdf5_18/1.8.21 gcc9/9.5.0 python39 \ cuda11.2/toolkit/11.2.2 cudnn8.1-cuda11.2/8.1.1.33 ml-pythondeps-py39-cuda11.2-gcc9/4.8.1 \ protobuf3-gcc9/3.9.2 nccl2-cuda11.2-gcc9/2.14.3 [root@basecm10 ~]# module load openmpi4-cuda11.2-ofed51-gcc9/ Loading openmpi4-cuda11.2-ofed51-gcc9/4.1.4 Loading requirement: hpcx/mlnx-ofed51/2.7.4 ucx/1.10.1 cm-pmix3/3.1.4 hwloc/1.11.11 [root@basecm10 ~]# python3.9 Python 3.9.16 (main, Mar 16 2023, 14:03:52) [GCC 11.2.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import tensorflow as tf ... >>> hello = tf.constant('TensorFlow 2 Hello World') ... 2023-03-22 17:20:10.514379: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created \ device /job:localhost/replica:0/task:0/device:GPU:0 with 31002 MB memory: -> device: 0, name: \ Tesla V100-SXM3-32GB, pci bus id: 0000:00:06.0, compute capability: 7.0 >>> ``` ```bash >>> hello = tf.constant('TensorFlow 2 Hello World') >>> tf.print(hello) TensorFlow 2 Hello World >>> a = tf.constant(10) >>> b = tf.constant(32) >>> tf.print(a+b) 42 >>> ``` ## 2.2 Deep Convolutional Generative Adversarial Network The following trains a deep convolutional generative adversarial network (DCGAN). The example code is included in the TensorFlow extras package. Once its module has been loaded, the example directory is defined with the `CM_TENSORFLOW2_EXTRA` environment variable. The example picks up training images and labels from the MNIST site, and places them in a directory `tensorflow_datasets/` if it needs to. The images are then used to train the model. End users would be expected to train the neural network within their home directories. **Example** ```bash [root@basecm10 ~]# module load tensorflow2-extra-py39-cuda11.2-gcc9 Loading tensorflow2-extra-py39-cuda11.2-gcc9/2.7.0 Loading requirement: openblas/dynamic/0.3.18 hdf5_18/1.8.21 gcc9/9.5.0 python39 \ cuda11.2/toolkit/11.2.2 cudnn8.1-cuda11.2/8.1.1.33 ml-pythondeps-py39-cuda11.2-gcc9/4.8.1\ protobuf3-gcc9/3.9.2 nccl2-cuda11.2-gcc9/2.14.3 tensorflow2-py39-cuda11.2-gcc9/2.7.0 opencv4-py39-cuda11.2-gcc9/4.5.4 [root@basecm10 ~]# module load openmpi4-cuda11.2-ofed51-gcc9/ Loading requirement: hpcx/mlnx-ofed51/2.7.4 ucx/1.10.1 cm-pmix3/3.1.4 hwloc/1.11.11 [root@basecm10 ~]# cd ${CM_TENSORFLOW2_EXTRA}/tensorflow_examples/models/dcgan/ [root@basecm10 dcgan]# python3.9 dcgan.py --epochs 5 ... I0322 18:24:37.097312 23456247932736 dataset_builder.py:400] Generating dataset mnist \ (/root/tensorflow_datasets/mnist/3.0.1) Downloading and preparing dataset Unknown size (download: Unknown size, generated: Unknown \ size, total: Unknown size) to /root/tensorflow_datasets/mnist/3.0.1... Dl Completed...: 0 url [00:00, ? url/s] I0322 18:24:37.484689 23456247932736 \ download_manager.py:354] Downloading https://storage.googleapis.com/cvdf-datasets/mnist/... ... Generating splits...: ... ... 2023-03-22 18:24:39.178613: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created \ device /job:localhost/replica:0/task:0/device:GPU:0 with 31002 MB memory: -> device: 0, name: \ Tesla V100-SXM3-32GB, pci bus id: 0000:00:06.0, compute capability: 7.0 ... Done writing mnist-train.tfrecord. Number of examples: 60000 (shards: [60000]) Generating splits...: ... Dataset mnist downloaded and prepared to /root/tensorflow_datasets/mnist/3.0.1. Subsequent calls will \ reuse this data. I0322 18:25:39.912966 23456247932736 logging_logger.py:35] Constructing tf.data.Dataset mnist for split \ train, from /root/tensorflow_datasets/mnist/3.0.1 Training ... 2023-03-22 18:25:42.881613: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8101 Epoch 0, Generator loss 1.0063459873199463, Discriminator Loss 1.0740859508514404 Epoch 1, Generator loss 0.9040709733963013, Discriminator Loss 1.0049669742584229 Epoch 2, Generator loss 0.88487708568573, Discriminator Loss 1.3763940334320068 Epoch 3, Generator loss 0.8470083475112915, Discriminator Loss 1.4865632057189941 Epoch 4, Generator loss 1.2192224264144897, Discriminator Loss 1.2187243700027466 [root@basecm10 dcgan]# ``` Datasets generated by the preceding run can be purged with: **Example** ```bash [root@basecm10 dcgan]# rm -rf /root/tensorflow_datasets/ ``` ## 2.3 Image-to-Image Translation with Conditional Adversarial Nets The following trains a conditional adversarial networks as a general-purpose solution to image-to-image translation problems. The trained model is capable of completing different tasks, such as colorizing black and white photos. The example code is included in the TensorFlow extras package. Once its module has been loaded, the example directory is defined with the `CM_TENSORFLOW2_EXTRA` environment variable. The example uses a preprocessed copy of the CMP Facade Database, helpfully provided by the Center for Machine Perception at the Czech Technical University in Prague. The original dataset includes 606 rectified images of facades from various sources, which have been manually annotated. The facades are from different cities around the world and diverse architectural styles. The example conveniently downloads the dataset in a temporary directory and then trains the model. End users would be expected to train the neural network within their home directories. **Example** ```bash [root@basecm10 ~]# module load tensorflow2-extra-py39-cuda11.2-gcc9 [root@basecm10 ~]# module load openmpi4-cuda11.2-ofed51-gcc9/ [root@basecm10 ~]# cd ${CM_TENSORFLOW2_EXTRA}/tensorflow_examples/models/pix2pix [root@basecm10 pix2pix]# python3.9 data_download.py --download_path /tmp ... Downloading data from http://efrosgans.eecs.berkeley.edu/pix2pix/datasets/facades.tar.gz ... [root@basecm10 pix2pix]# python3.9 pix2pix.py --path /tmp/facades --epochs 5 ... 2023-03-22 20:11:12.173915: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created \ device /job:localhost/replica:0/task:0/device:GPU:0 with 31002 MB memory: -> device: 0,\ name: Tesla V100-SXM3-32GB, pci bus id: 0000:00:06.0, compute capability: 7.0 Training ... 2023-03-22 20:11:16.660300: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8101 Epoch 0, Generator loss 39.03266906738281, Discriminator Loss 0.23912914097309113 Epoch 1, Generator loss 41.69317626953125, Discriminator Loss 0.07581206411123276 Epoch 2, Generator loss 33.53696823120117, Discriminator Loss 0.13100799918174744 Epoch 3, Generator loss 30.571992874145508, Discriminator Loss 1.0946425199508667 Epoch 4, Generator loss 29.1583309173584, Discriminator Loss 1.0569590330123901 [root@basecm10 pix2pix]# ``` ## 2.4 Neural Machine Translation With Attention The following trains a sequence to sequence model for neural machine translation with attention using gated recurrent units (GRUs). The example code is included in the TensorFlow extra package. Once its module has been loaded, the example directory is defined with the `CM_TENSORFLOW2_EXTRA` environment variable. The example conveniently downloads a dataset for Spanish to English translation in a temporary directory and then trains the model. End users would be expected to train the neural network within their home directories. **Example** ```bash [root@basecm10 ~]# module load tensorflow2-extra-py39-cuda11.2-gcc9 [root@basecm10 ~]# module load openmpi4-cuda11.2-ofed51-gcc9 [root@basecm10 ~]# cd ${CM_TENSORFLOW2_EXTRA}/tensorflow_examples/models/nmt_with_attention/ [root@basecm10 nmt_with_attention]# python3.9 train.py --epochs 5 --download_path /tmp ... Downloading data from http://storage.googleapis.com/download.tensorflow.org/data/spa-eng.zip ... 2023-03-22 21:18:11.717144: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created \ device /job:localhost/replica:0/task:0/device:GPU:0 with 31002 MB memory: -> device: 0,\ name: Tesla V100-SXM3-32GB, pci bus id: 0000:00:06.0, compute capability: 7.0 Training ... 2023-03-22 21:18:27.574448: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8101 Epoch: 0, Train Loss: 1.814919114112854, Test Loss: 2.885890483856201 Epoch: 1, Train Loss: 1.085754156112671, Test Loss: 2.093322515487671 Epoch: 2, Train Loss: 0.6579596996307373, Test Loss: 1.7698687314987183 Epoch: 3, Train Loss: 0.43058785796165466, Test Loss: 1.7104995250701904 Epoch: 4, Train Loss: 0.29615747928619385, Test Loss: 1.7149256467819214 [root@basecm10 nmt_with_attention]# ```