Develop Reference

How to run an application based on TensorFlow 2 in Docker container?

I am relatively new to TensorFlow, so I have been trying to run simple applications locally, and everything was going well.
At some point I wanted to Dockerize my application. Building the Docker image went with no errors, however, when I tried to run my application, I received the following error:
AttributeError: module 'tensorflow' has no attribute 'gfile'. Did you mean: 'fill'?
After googling about the problem, I understood that it is caused by version differences between TF1 and TF2.
One of the explanation about the problem I found is found here.
Locally, I am using TF2 (specifically 2.9.1), inside a virtual environment.
When dockerizing, I also confirmed from inside the docker container that my TF version is the same.
I also tried to run the container in interactive mode, and create virtual environment, and install all dependencies manually, exactly the same way I did locally, but still with no success.
My Dockerfile is as follows:
FROM python:3-slim
# ENV VIRTUAL_ENV=/opt/venv
# RUN python3 -m venv $VIRTUAL_ENV
# ENV PATH="$VIRTUAL_ENV/bin:$PATH"
WORKDIR /objectDetector
RUN apt-get update
RUN apt-get install -y protobuf-compiler
RUN apt-get install ffmpeg libsm6 libxext6 -y
RUN pip3 install update && python3 -m pip install --upgrade pip
RUN pip3 install tensorflow==2.9.1
RUN pip3 install tensorflow-object-detection-api
RUN pip3 install opencv-python
RUN pip3 install opencv-contrib-python
COPY detect_objects.py .
COPY detector.py .
COPY helloWorld.py .
ADD data data /objectDetector/data/
ADD models /objectDetector/models/
So my question is: How can I ran an application using TensorFlow 2 from a docker container?
Am I missing something here?
Thanks in advance for any help or explanation.

I believe that in tensorflow 2.0 :
tf.gfile was replaced by tf.io.gfile
Can you try this ?
Have a nice day,
Gabriel

How do I install dateinfer inside my Docker image

Some background : I'm new to understanding docker images and containers and how to write DOCKERFILE. I currently have a Dockerfile which installs all the dependencies that I want through PIP install command and so, it was very simple to build and deploy images.
But I currently have a new requirement to use the Dateinfer module and that cannot be installed through the pip install command.
The repo has to be first cloned and then has to be installed and I'm having difficulty achieving this through a DOCKERFILE. The current work around I've been following for now is to run the container and install it manually in the directory with all the other dependencies and Committing the changes with dateinfer installed.But this is a very tedious and time consuming process and I want to achieve the same by just mentioning it in the DOCKERFILE along with all my other dependencies.
This is what my Dockerfile looks like:
FROM ubuntu:20.04
RUN apt update
RUN apt upgrade -y
RUN apt-get install -y python3
RUN apt-get install -y python3-pip
RUN DEBIAN_FRONTEND=noninteractive TZ=Etc/UTC apt-get -y install tzdata
RUN apt-get install -y libenchant1c2a
RUN apt install git -y
RUN pip3 install argparse
RUN pip3 install boto3
RUN pip3 install numpy==1.19.1
RUN pip3 install scipy
RUN pip3 install pandas
RUN pip3 install scikit-learn
RUN pip3 install matplotlib
RUN pip3 install plotly
RUN pip3 install kaleido
RUN pip3 install fpdf
RUN pip3 install regex
RUN pip3 install pyenchant
RUN pip3 install openpyxl
ADD core.py /
ENTRYPOINT [ "/usr/bin/python3.8", "/core.py”]
So when I try to install Dateinfer like this:
RUN git clone https://github.com/nedap/dateinfer.git
RUN cd dateinfer
RUN pip3 install .
It throws the following error :
ERROR: Directory '.' is not installable. Neither 'setup.py' nor 'pyproject.toml' found.
The command '/bin/sh -c pip3 install .' returned a non-zero code: 1
How do I solve this?

Each RUN directive in a Dockerfile runs in its own subshell. If you write something like this:
RUN cd dateinfer
That is a no-op: it starts a new shell, changes directory, and then the shell exits. When the next RUN command executes, you're back in the / directory.
The easiest way of resolving this is to include your commands in a single RUN statement:
RUN git clone https://github.com/nedap/dateinfer.git && \
cd dateinfer && \
pip3 install .
In fact, you would benefit from doing this with your other pip install commands as well; rather than a bunch of individual RUN
commands, consider instead:
RUN pip3 install \
argparse \
boto3 \
numpy==1.19.1 \
scipy \
pandas \
scikit-learn \
matplotlib \
plotly \
kaleido \
fpdf \
regex \
pyenchant \
openpyxl
That will generally be faster because pip only needs to resolve
dependencies once.
Rather than specifying all the packages individually on the command
line, you could also put them into a requirements.txt file, and then
use pip install -r requirements.txt.

Python- Unable to Train Tensorflow Model Container in Sagemaker

I'm fairly new to Sagemaker and Docker.I am trying to train my own custom object detection algorithm in Sagemaker using an ECS container. I'm using this repo's files:
https://github.com/svpino/tensorflow-object-detection-sagemaker
I've followed the instructions exactly, and I'm able to run the image in a container perfectly fine on my local machine. But when I push the image to ECS to run in Sagemaker, I get the following message in Cloudwatch:
I understand that for some reason, when deployed to ECS suddenly the image can't find python. At the top of my training script is the text #!/usr/bin/env python. I've tried to run the *which python * command and changed up text to point to #!/usr/local/bin python, but I just get additional errors. I don't understand why this image would work on my local (tested with both docker on windows and docker CE for WSL). Here's a snippet of the docker file:
ARG ARCHITECTURE=1.15.0-gpu
FROM tensorflow/tensorflow:${ARCHITECTURE}-py3
RUN apt-get update && apt-get install -y --no-install-recommends \
wget zip unzip git ca-certificates curl nginx python
# We need to install Protocol Buffers (Protobuf). Protobuf is Google's language and platform-neutral,
# extensible mechanism for serializing structured data. To make sure you are using the most updated code,
# replace the linked release below with the latest version available on the Git repository.
RUN curl -OL https://github.com/protocolbuffers/protobuf/releases/download/v3.10.1/protoc-3.10.1-linux-x86_64.zip
RUN unzip protoc-3.10.1-linux-x86_64.zip -d protoc3
RUN mv protoc3/bin/* /usr/local/bin/
RUN mv protoc3/include/* /usr/local/include/
# Let's add the folder that we are going to be using to install all of our machine learning-related code
# to the PATH. This is the folder used by SageMaker to find and run our code.
ENV PATH="/opt/ml/code:${PATH}"
RUN mkdir -p /opt/ml/code
WORKDIR /opt/ml/code
RUN pip install --upgrade pip
RUN pip install cython
RUN pip install contextlib2
RUN pip install pillow
RUN pip install lxml
RUN pip install matplotlib
RUN pip install flask
RUN pip install gevent
RUN pip install gunicorn
RUN pip install pycocotools
# Let's now download Tensorflow from the official Git repository and install Tensorflow Slim from
# its folder.
RUN git clone https://github.com/tensorflow/models/ tensorflow-models
RUN pip install -e tensorflow-models/research/slim
# We can now install the Object Detection API, also part of the Tensorflow repository. We are going to change
# the working directory for a minute so we can do this easily.
WORKDIR /opt/ml/code/tensorflow-models/research
RUN protoc object_detection/protos/*.proto --python_out=.
RUN python setup.py build
RUN python setup.py install
# If you are interested in using COCO evaluation metrics, you can tun the following commands to add the
# necessary resources to your Tensorflow installation.
RUN git clone https://github.com/cocodataset/cocoapi.git
WORKDIR /opt/ml/code/tensorflow-models/research/cocoapi/PythonAPI
RUN make
RUN cp -r pycocotools /opt/ml/code/tensorflow-models/research/
# Let's put the working directory back to where it needs to be, copy all of our code, and update the PYTHONPATH
# to include the newly installed Tensorflow libraries.
WORKDIR /opt/ml/code
COPY /code /opt/ml/code
ENV PYTHONPATH=${PYTHONPATH}:tensorflow-models/research:tensorflow-models/research/slim:tensorflow-models/research/object_detection
RUN chmod +x /opt/ml/code/train
CMD ["/bin/bash","-c","chmod +x /opt/ml/code/train && /opt/ml/code/train"]

Not able to install any python package in docker container

I am trying to create an docker image with ubutu 16.04 as base. I want to install few python packages like pandas, flask etc. I have kept all packages in "requirements.txt". But when I am trying to build image, I am getting
Could not find a version that satisfies the requirement requests (from -r requirements.txt (line 1)) (from versions: )
No matching distribution found for requests (from -r requirements.txt (line 1))
Basically, I have not mentioned any version in "requirements.txt". I guess it should take the latest available and compatible version of that package. But for every package same issue I am getting.
My DockerFile is as follows.
FROM ubuntu:16.04
RUN apt-get update -y && \
apt-get install -y python3-pip python3-dev build-essential cmake pkg-config libx11-dev libatlas-base-dev
# We copy just the requirements.txt first to leverage Docker cache
COPY ./requirements.txt /testing/requirements.txt
WORKDIR /testing
RUN pip3 install -r requirements.txt
and requirements.txt is.
pandas
requests
PyMySQL
Flask
Flask-Cors
Pillow
face-recognition
Flask-SocketIO
Where I am doing wrong ? Can anybody help ?

I too ran into the same situation. I observed that, python packages is looking for the network within docker. It is thinking that, it is running in a standalone without network so its not able to locate the package. In these type of situations either
No matching distribution found
or sometimes
Retrying ...
error may occur.
I used a --network option in the docker build command like below to overcome this error where the command insists python to use the host network to download the required packages.
docker build --network=host -t tracker:latest .

Try using this:
RUN python3.6 -m pip install --upgrade pip \
&& python3.6 -m pip install -r requirements.txt
by using it in this way, you are specifying the version of python in which you want to search for those packages.
Change it to python3.7 if you wish to use 3.7 version.

I suggest using the official python image instead. As a result, your Dockerfile will now become:
FROM python:3
WORKDIR /testing
COPY ./requirements.txt /testing/requirements.txt
RUN pip install --no-cache-dir -r requirements.txt
... etc ...
Now re: Angular/Node. You have two options from here: 1) Install Angular/Node on the Python image; or 2) Use Docker's multi-stage build feature so you build the Angular and Python-specific images before merging them together. Option 2 is recommended but it would take some work. It would probably look like this:
FROM node:8 as node
# Angular-specific build
FROM python:3 as python
# Python-specific build
# Then copy your data from the Angular image to the Python one:
COPY --from=node /usr/src/app/dist/angular-docker /usr/src/app

How to run SSL_library_init() from Python 3.7 docker image

Up until recently I've been using openssl library within python:3.6.6-jessie docker image and thing worked as intented.
I'm using very basic Dockerfile configuration to install all necessary dependencies:
FROM python:3.6.6-jessie
ENV PYTHONDONTWRITEBYTECODE 1
ENV PYTHONUNBUFFERED 1
WORKDIR /code
RUN apt-get -qq update
RUN apt-get install openssl
RUN apt-get upgrade -y openssl
ADD requirements.txt /code/
RUN pip install --upgrade pip
RUN pip install -r requirements.txt
And access and initialize the library itself with these 2 lines:
openssl = cdll.LoadLibrary("libssl.so")
openssl.SSL_library_init()
Things were working great with this approach.
This week I was doing upgrade of python and libraries and as result I switched to newer docker image:
FROM python:3.7.5
...
This immediatelly caused openssl to stop working because of this exception:
AttributeError: /usr/lib/x86_64-linux-gnu/libssl.so.1.1: undefined symbol: SSL_library_init
From this error I can understand that libssl no longer provides SSL_library_init method (or so it seems to be) which is rather weird issue because the initializer name in openssl documentation is the same.
I also tried to resolve this using -stretch and -buster distributions but the issue remains.
What is the correct approach to run SSL_library_init in those newer distributions? Maybe some additional dockerfile configuration is required?

I think you need to install libssl1.0-dev
RUN apt-get install -y libssl1.0-dev