[Task]: Build docker image for tensor RT tests during Run time

[AnandInguva]

What needs to happen?

Build the docker container with all the necessary dependencies for tensorRT frameworks during run time. This would make the testing easier if there is any change in the dependencies such that container can be build with those changes and can be tested on the code.

Issue Priority

Priority: 2

Issue Component

Component: testing

[AnandInguva]

.take-issue

[AnandInguva]

.add-labels run-inference

[yeandy]

Used dockerfile with the contents

ARG BUILD_IMAGE=nvcr.io/nvidia/tensorrt:22.05-py3

FROM ${BUILD_IMAGE} 

ENV PATH="/usr/src/tensorrt/bin:${PATH}"

WORKDIR /workspace

RUN pip install --no-cache-dir apache-beam[gcp]==2.40.0
COPY --from=apache/beam_python3.8_sdk:2.40.0 /opt/apache/beam /opt/apache/beam

RUN pip install --upgrade pip \
    && pip install torch>=1.7.1 \
    && pip install torchvision>=0.8.2 \
    && pip install pillow>=8.0.0 \
    && pip install transformers>=4.18.0 \
    && pip install cuda-python

ENTRYPOINT [ "/opt/apache/beam/boot" ]

[damccorm]

@AnandInguva is there more to do here or should we close this one?

[AnandInguva]

It is a tracking issue for building a docker image during runtime for tensorRT. There is some more work to do here

[Abacn]

The latest container is half year old now, and now it takes ~15 minutes to resolve the dependency when installing the Beam SDK for :sdks:python:test-suites:dataflow:py38:tensorRTtests running on Python3.8 PostCommit. Update Python 3.8 container may also be considered in the scope of this issue