tensorflow

tensorflow

CONTAINERS IMAGES RUN BUILD

Legacy TF1 installer and TF2 builder

CONTAINERS

tensorflow2:2.16.1
Requires	L4T ['>=36']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn:9.3 python tensorrt pybind11 numpy h5py bazel protobuf:cpp
Dockerfile	Dockerfile
Notes	TensorFlow TF2 version 2.16.1

tensorflow2:2.18.0
Aliases	tensorflow2:2.18.0 tensorflow tensorflow2
Requires	L4T ['>=36']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn:9.3 python tensorrt pybind11 numpy h5py bazel protobuf:cpp
Dependants	l4t-ml l4t-tensorflow:tf1 l4t-tensorflow:tf2 tensorflow_graphics:2.18.0 tensorflow_text:2.18.0
Dockerfile	Dockerfile.pip
Notes	TensorFlow TF2 version 2.18.0 (will be built from source)

tensorflow2:2.19.0
Aliases	tensorflow2:2.19.0
Requires	L4T ['>=36']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn:9.3 python tensorrt pybind11 numpy h5py bazel protobuf:cpp
Dependants	tensorflow_graphics:2.19.0 tensorflow_text:2.19.0
Dockerfile	Dockerfile.pip
Notes	TensorFlow TF2 version 2.19.0 (will be built from source)

tensorflow2:2.20.0
Aliases	tensorflow2:2.20.0
Requires	L4T ['>=36']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn:9.3 python tensorrt pybind11 numpy h5py bazel protobuf:cpp
Dependants	tensorflow_graphics:2.20.0 tensorflow_text:2.20.0
Dockerfile	Dockerfile.pip
Notes	TensorFlow TF2 version 2.20.0 (will be built from source)

tensorflow2:2.21.0
Aliases	tensorflow2:2.21.0
Requires	L4T ['>=36']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn:9.3 python tensorrt pybind11 numpy h5py bazel protobuf:cpp
Dockerfile	Dockerfile.pip
Notes	TensorFlow TF2 version 2.21.0 (will be built from source)

CONTAINER IMAGES

Repository/Tag	Date	Arch	Size
dustynv/tensorflow:r32.7.1	2023-12-06	arm64	0.8GB
dustynv/tensorflow:r35.2.1	2023-12-05	arm64	5.5GB
dustynv/tensorflow:r35.3.1	2023-08-29	arm64	5.5GB
dustynv/tensorflow:r35.4.1	2023-12-06	arm64	5.5GB

_{Container images are compatible with other minor versions of JetPack/L4T:
    • L4T R32.7 containers can run on other versions of L4T R32.7 (JetPack 4.6+)
    • L4T R35.x containers can run on other versions of L4T R35.x (JetPack 5.1+)}

RUN CONTAINER

To start the container, you can use jetson-containers run and autotag, or manually put together a docker run command:

# automatically pull or build a compatible container image
jetson-containers run $(autotag tensorflow)

# or explicitly specify one of the container images above
jetson-containers run dustynv/tensorflow:r32.7.1

# or if using 'docker run' (specify image and mounts/ect)
sudo docker run --runtime nvidia -it --rm --network=host dustynv/tensorflow:r32.7.1

^{jetson-containers run forwards arguments to docker run with some defaults added (like --runtime nvidia, mounts a /data cache, and detects devices)
autotag finds a container image that's compatible with your version of JetPack/L4T - either locally, pulled from a registry, or by building it.}

To mount your own directories into the container, use the -v or --volume flags:

jetson-containers run -v /path/on/host:/path/in/container $(autotag tensorflow)

To launch the container running a command, as opposed to an interactive shell:

jetson-containers run $(autotag tensorflow) my_app --abc xyz

You can pass any options to it that you would to docker run, and it'll print out the full command that it constructs before executing it.

BUILD CONTAINER

If you use autotag as shown above, it'll ask to build the container for you if needed. To manually build it, first do the system setup, then run:

jetson-containers build tensorflow

The dependencies from above will be built into the container, and it'll be tested during. Run it with --help for build options.