Valentin

+1 Not everyone runs the very modern V100s. Therefore looking into results with K80 and P100 were still interesting from a practical point-of-view. It's just unfair that the results are...

Thank you, @derekhh , this is great! Do you also know, if /where to find corresponding plots? and if the README.md at [1] will be changed accordingly ("Benchmark results can...

Hi @mtreinish , Thank you very much for testing this. My configuration can actually work if I change ``.stestr.conf`` as: ``` [DEFAULT] test_path=./src/tests/ top_dir=./ ``` i.e instead of ``top_dir=./src`` specify...

@IgnacioHeredia Could you, please, elaborate more? What is the exact cli call and what line is suspicious? Because predict() supposes to return json response not "success". "return 0" is anyway...

Hi @IgnacioHeredia thanks for updating, I will have a look, though need some time to refresh my memory :-) One quick comment: if I recon correctly, "fields.Field" is a general...

In a very general case, data-processing might be an asset in itself, like we may have modules which do only 'training' (e.g. benchmarks), we may have modules with only predict...

@ericof , @kurtmckee , after... some time a colleague of mine @BorjaEst added requested tests. One CI/CD test (3.7) is failing but the complain is about vulnerability in Jinja2 version...