Adding Decision Forests Op #12
Comments
|
Thanks for the note of appreciation! Additional tensorflow ops for tensorflow-serving-arm is something I have not yet considered, so it's very possible it's just not supported in this project yet. I'll have a look over those instructions and with a little luck hopefully it will be straight forward to implement support here. As for the images, if I am building all arches (which I don't do too often), I tend to just run something like: bazel run //tensorflow_model_server:project_image --config=release --config=linux_amd64_avx_sse4.2 && \
bazel run //tensorflow_model_server:project_image_debug --config=release --config=linux_amd64_avx_sse4.2 && \
bazel run //tensorflow_model_server:project_image --config=release --config=linux_arm64_armv8-a && \
bazel run //tensorflow_model_server:project_image_debug --config=release --config=linux_arm64_armv8-a && \
bazel run //tensorflow_model_server:project_image --config=release --config=linux_arm64_armv8.2-a && \
bazel run //tensorflow_model_server:project_image_debug --config=release --config=linux_arm64_armv8.2-a && \
bazel run //tensorflow_model_server:project_image --config=release --config=linux_arm_armv7-a_neon_vfpv4 && \
bazel run //tensorflow_model_server:project_image_debug --config=release --config=linux_arm_armv7-a_neon_vfpv4 && \
bazel run //tensorflow_model_server:project_image --config=release --config=linux_arm_armv7-a_neon_vfpv3 && \
bazel run //tensorflow_model_server:project_image_debug --config=release --config=linux_arm_armv7-a_neon_vfpv3publishing (multi-arch manifest creation) is scripted in: |
|
ok, one thing to try before digging deeper: follow all proposed instructions except for this modification: instead of changing
|
|
Thank you for the reply |
|
Ok, you were spot on- adding that into that build file cause the df stuff to be build, but it seems to error: This is what I have so far. I am then trying to compile it with: My suspicion is that the Appologies for documenting this here and I don't expect this repo to support this in the end, but more than willing to add any findings to a readme in the end. |
|
can you post the compile command that generates the above error? |
|
This one? |
|
I just came across this issue - @grantstephens I'm not sure if you ever got this working but Tensorflow serving, as of version 2.11.0, natively supports Tensorflow Decision Forests. @emacski Any chance you fancy making some new Docker images based upon Tensorflow serving 2.11.0 so we can get Decision Forests on M1? :) I am not yet on M1 but have colleagues who need this! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi- thank you so much for this repo- it has been a lifesaver so far.
I am trying to follow these instructions whilst still compiling for arm, but running into some trouble.
I added the needed lines to a new workspace file inf the tensorflow_model_server folder and added the
SUPPORTED_TENSORFLOW_OPSto the BUILD file in that folder. It compiled ok and I got a docker image out of it, but the added ops aren't present.Any help would be much appreciated.
Oh, one last thing- is there a command to compile an image for all architectures, or do you have to manually combine them all at the end?
The text was updated successfully, but these errors were encountered: