build: change custom build scripts to build production image

[rstol]

No description provided.

[JeffWigger]

I have changed the scripts such that they install the version of PyTorch that we actually want and tested it on DEV.
I will wait with merge until the torch serve PR is merged.

[pypae]

I just built the cpu image on my Mac and pushed it to docker hub.

I also made the pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu tag multi-platform following the docker manifest docs:

Here's the commands I used:

# Build the arm image and tag it:
./build_custom_images.sh
docker tag textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-v2-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm
docker push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm

# Pull the amd image and tag it:
docker pull textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu
docker tag textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd
docker push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd

# Create a multi-platform manifest
docker manifest create textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu \
    textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd \
    textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm

# Annotate the arm version
docker manifest annotate textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm --arch arm64

# Push the manifest
docker manifest push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu

[mpoemsl]

Note that I switched back to using the Dockerfile.dev since Dockerfile was just installing the factory torchserve version from pip.

[mpoemsl]

I just built the cpu image on my Mac and pushed it to docker hub.

I also made the pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu tag multi-platform following the docker manifest docs:

Here's the commands I used:

# Build the arm image and tag it:
./build_custom_images.sh
docker tag textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-v2-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm
docker push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm

# Pull the amd image and tag it:
docker pull textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu
docker tag textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd
docker push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd

# Create a multi-platform manifest
docker manifest create textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu \
    textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-amd \
    textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm

# Annotate the arm version
docker manifest annotate textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu-arm --arch arm64

# Push the manifest
docker manifest push textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-pt21-cpu

I have now created a multi-arch image textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-v3-cpu and a single-arch amd image textshuttle/pytorch-serve:torchserve-23mt-v0.8.0-v3-gpu from #16.