Add test data

[loostrum]

I've removed the test data instructions from the README and added a .tar.gz file containing both the model and test data. The test data can be verified by running python workflow_scripts/test_models.py --target onnxruntime in the root of the repo. It should output something like this:

==============Testing leafsnap30.tar.gz==============
LFS pull completed for vision/classification/leafsnap30/model/leafsnap30.tar.gz with return code= 0
Running tests in /Users/loostrum/projects/DIANNA/models/ci_test_tar_dir/leafsnap30 for /Users/loostrum/projects/DIANNA/models/ci_test_tar_dir/leafsnap30/leafsnap_model.onnx
/Users/loostrum/projects/DIANNA/models/ci_test_tar_dir/leafsnap30/test_data_set_0
PASS
[PASS] leafsnap30.tar.gz is checked by onnxruntime. 
LFS prune completed with return code= 0
==============Testing leafsnap_model.onnx==============
LFS prune completed with return code= 0
2 models have been checked. 

There is one more step that I can't get to work, perhaps someone else can help out here. The instructions are the following:
If this PR does update/add .onnx or .tar.gz files, please use python workflow_scripts/generate_onnx_hub_manifest.py --target diff to update ONNX_HUB_MANIFEST.json with according model information (especially SHA) for ONNX Hub.

However, when I run the script the json file is not updated. The script doest detected the leafsnap model though.

[loostrum]

I've updated the .json file manually by copying from the MNIST entry and updating the values. Ready for review now. To review the values in the json file: the shasum and bytes fields must match those given by github lfs for those files.

[elboyran]

Looks good. Will submit to ONNX and see what happens.