search

dusty-nv / jetson-containers

Machine Learning Containers for NVIDIA Jetson and JetPack-L4T
MIT License
1.93k stars 421 forks source link

./scripts/docker_build_ml.sh all # build all: l4t-pytorch, l4t-tensorflow, and l4t-ml #63

Open abdo-diwan opened 3 years ago

abdo-diwan commented 3 years ago

done testing container l4t-tensorflow:r32.5.1-tf2.3-py3 => numpy

testing container l4t-ml:r32.5.1-py3 => PyTorch xhost: unable to open display "" Unable to find image 'l4t-ml:r32.5.1-py3' locally docker: Error response from daemon: pull access denied for l4t-ml, repository does not exist or may require 'docker l ogin': denied: requested access to the resource is denied. See 'docker run --help'.

abdo-diwan commented 3 years ago

Hi, I have jetson-nano-jp451 and I am trying to build all:

I just modified the Dockerfile.ml by adding this

RUN pip3 install matplotlib --verbose RUN pip3 install seaborn--verbose RUN pip3 install psutil --verbose RUN pip3 install GPUtil --verbose

./scripts/docker_build_ml.sh all # build all: l4t-pytorch, l4t-tensorflow, and l4t-ml

but when i check docker images i have ( l4t-tensorflow, l4t-pytorch ) i do not have l4t-ml image

Successfully built 2ec9f823705c Successfully tagged l4t-tensorflow:r32.5.1-tf2.3-py3 done building TensorFlow tensorflow-2.3.1+nv20.11-cp36-cp36m-linux_aarch64.whl, l4t-tensorflow:r32.5.1-tf2.3-py3 Building l4t-ml:r32.5.1-py3 container... Sending build context to Docker daemon 238.6kB Step 1/40 : ARG BASE_IMAGE=nvcr.io/nvidia/l4t-base:r32.4.4 Step 2/40 : ARG PYTORCH_IMAGE Step 3/40 : ARG TENSORFLOW_IMAGE Step 4/40 : FROM ${PYTORCH_IMAGE} as pytorch pull access denied for l4t-pytorch, repository does not exist or may require 'docker login': denied: requested access to the resource is denied

diwan@diwan-desktop:/jetson-containers$ sudo docker images REPOSITORY TAG IMAGE ID CREATED SIZE l4t-tensorflow r32.5.1-tf2.3-py3 2ec9f823705c 7 minutes ago 2.05GB l4t-tensorflow r32.5.1-tf1.15-py3 41e80e990bcf 42 minutes ago 2.12GB l4t-pytorch r32.5.1-pth1.8-py3 663e903d6d14 2 hours ago 1.86GB nvcr.io/nvidia/l4t-base r32.5.0 d9141f00a2c3 4 months ago 635MB diwan@diwan-desktop:/jetson-containers$

dusty-nv commented 3 years ago

Hi @abdo-diwan, try changing pth-1.7 from this line of the build script to pth-1.8 instead:

https://github.com/dusty-nv/jetson-containers/blob/1e10908a104494a883f6855d1e9947827f2a17bc/scripts/docker_build_ml.sh#L131

abdo-diwan commented 3 years ago

@dusty-nv

installed successfully and docker is running also jupyter lab is working but when i try to lunch the classic jupyter notebook and open .ipynb files i still got an error "500 : Internal Server Error" do you know what shold i do to avoid that ?

the idea that i need to use notebook backend, which is fully compatible with matplotlib-FunAnimation

dusty-nv commented 3 years ago

Hi @abdo-diwan, I haven't tested the classic jupyter before, just jupyterlab. jupyterlab can open ipynb notebooks as well.