Torch handler Unit test : AssertionError for test_object_detector.py

[abishekchiffon]

Context

The handler unit tests in the ts/torch_handler folder fails in the master branch.
As you run "./run_unit_test.sh" , it fails with the following error

AssertionError: assert 'bench' in [{'person': [167.4222869873047, 57.03825378417969, 301.305419921875, 436.68682861328125], 'score': 0.9995299577713013}...bag': [379.4246826171875, 259.97760009765625, 419.0148620605469, 317.9510498046875], 'score': 0.9896279573440552}, ...]

in the test_object_detector.py file in the line assert('bench' in results[0]).

• torchserve version:0.2.0b20201104
• torch version:1.6.0
• torchvision version [if any]:0.7.0
• torchtext version [if any]:0.7.0
• torchaudio version [if any]: NA
• java version:11.0.5
• Operating System and version: ubuntu 16.04

Your Environment

• Installed using source? [yes/no]: yes
• Are you planning to deploy it using docker container? [yes/no]: no
• Is it a CPU or GPU environment?: CPU
• Using a default/custom handler? [If possible upload/share custom handler/model]: NA
• What kind of model is it e.g. vision, text, audio?: NA
• Are you planning to use local models from model-store or public url being used e.g. from S3 bucket etc.? local models
  [If public url then provide link.]:
• Provide config.properties, logs [ts.log] and parameters used for model registration/update APIs: NA
• Link to your project [if any]: NA

Expected Behavior

The object detector unit test should pass.

Current Behavior

AssertionError: assert 'bench' in [{'person': [167.4222869873047, 57.03825378417969, 301.305419921875, 436.68682861328125], 'score': 0.9995299577713013}...bag': [379.4246826171875, 259.97760009765625, 419.0148620605469, 317.9510498046875], 'score': 0.9896279573440552}, ...]

Possible Solution

Modify the object_detector_unit_test to properly asserts the object_detector response.

Steps to Reproduce

1. Run ./scripts/install_from_src
   2.Run "run_unit_test.sh" inside the folder ts/torch_handler/unit_tests

Failure Logs [if any]

unit_test_metrics_object_detector.log