Vikas-kum

> Hi, I am unable to run that shell script. I am trying a blog https://aws.amazon.com/blogs/machine-learning/pruning-machine-learning-models-with-amazon-sagemaker-debugger-and-amazon-sagemaker-experiments/ and I got this error on line "from smdebug.trials import create_trial": [image: image.png] So...

I can't see the image, It's not public. Have you tried changing Line 16 and 17 to run with sudo . Also try this : https://stackoverflow.com/questions/35896335/how-can-i-uninstall-protobuf-3-0-0 btw what machine(macOSx, linux)...

This log is repeated because if range steps is defined, s3 list calls are still made but the contents are filtered. Can we not make list calls if range steps...

I have done some optimization. Current time is 38 minutes for tests to run. Some unit tests are taking more than 5 minutes. Reports can be found in CI logs....

Time taken by unit test needs to be optimized 23.35s call tests/analysis/trials/test_s3.py::test_s3 22.24s call tests/analysis/trials/test_s3.py::test_multiple_s3_trials 9.83s call tests/analysis/trials/test_refresh.py::test_refresh_with_range_s3 8.48s call tests/analysis/trials/test_refresh.py::test_refresh_s3 7.83s call tests/analysis/trials/test_refresh.py::test_no_refresh_s3 5.14s call tests/analysis/tensors/test_refresh.py::test_refresh_tensors 4.85s call tests/analysis/trials/test_create.py::test_creation_s3...

@NihalHarish Please make the changes that we discussed last week offline.

can you check mode and mode step of saved global steps?

Enable CI pipeline with new docker TF 2.2 image.

Currently we dont have SageMaker Debugger support with HPO. But if you wish to use smdebug and has training running in internet mode, you can create smdebug hook in training...

Hi @austinmw . Thanks for reporting this. Can you please provide the example script to reproduce this.