1. Introduce
It is the Linux AI module test framework basing on pytest.

2. Folder
├── data
├── output
├── README
├── tests
└── utils
2.1 - data
    It includes case configuration, building configuration, model_zoo, dataset.
    ├── build_config
    │   └── zhouyi
    ├── case_config
    ├── dataset -> /swtest/NPU
    └── model_zoo -> /home/anthony/workspace/git/cix/npu_model_zoo
    The dataset and model_zoo are link to release dataset folder and model_zoo git.
    User needs set it in PC as an procedure of environment preparing.

2.2 - output
    It includes the intermediate result and final result of case running.
    Results of phase1(pre-process), phase2(DUT inference) and phase3(post-process) are stored
    under output with sub-folder named with case name.
    The final physical store place of output is the PC running phase1 and phase3.
    When DUT runs phase2, it needs mount this framework in board to access result of phase1 and
    store its result back for phase3.
    To speed up the inference process, in phase2, DUT need to copy the output to local storage.
    Also copying phase2 output back to nfs is needed.

2.3 - tests
    Includes the test cases scripts.

2.4 - utils
    Include the tools used in test cases.

3. How to run it
Example:
python -m pytest -k 'Classify-SM-0301' | tee 0618_sm0301.log

4. How to expand it
If the model type is already supported like classification models:
    Add building configuration under data/build_config folder
    Add case configuration item to data/case_config/xxx files
    Try to run and tuning the cases
For model type is not in the supporting list, below steps are needed:
    Add input pre_process steps
    Add inference result post_process steps
