Where is `list_topologies.yml`?

[banderlog]

list_topologies.yml dissapeared from downloader dir. downloader.py does not work without it.

python3 ./downloader.py --print_all
python3 ./downloader.py --name face-detection-adas-0001
>> No matching models: "face-detection-adas-0001"

What happened?

[vladimir-dudnik]

please check OpenVINO 2019 R3 Release notes

[banderlog]

Moved Model Downloader tool configuration files to separate per-model folders in order to improve user experience and simplify contribution process (less merge conflicts while developing/merging several models at the same time). The list is extended to support the following public models in Caffe, TensorFlow, MXNet, and PyTorch* formats:

And where are they?

This component includes documentation for the latest stable set of pre-trained models from https://github.com/opencv/open_model_zoo. The models in a binary form can be downloaded using the Model Downloader tool.

And it has broken link, but it directs exactly where I looked.

Could someone be so kind to tell me how users supposed to download models in RC3, maybe with code example and directory listing?

[vladimir-dudnik]

Starting from 2019 R3 downloader config files located in model's folder, and called model.yml, what you may see on example for AlexNet. To download models I've used the same command as I did before: python3 downloader.py --all -o MY_OUTPUT_DIR

[banderlog]

[ranger]banderlog@oldbook:~/Downloads/downloader$ python3 ./downloader.py --all -o OUT_DIR
>>################|| Downloading models ||################
>>
>>################|| Post-processing ||################

[ranger]banderlog@oldbook:~/Downloads/downloader$ python3 ./downloader.py --list ./tests/representative-models.lst
>>No matching models: "mobilenet-v1-0.25-128"

[ranger]banderlog@oldbook:~/Downloads/downloader$ tree
.
├── common.py
├── converter.py
├── downloader.py
├── info_dumper.py
├── license.txt
├── __pycache__
│   └── common.cpython-36.pyc
├── pytorch_to_onnx.py
├── README.md
├── requirements.in
├── requirements-pytorch.in
└── tests
    └── representative-models.lst

How it supposed to load all models if it has no files from where it can parse download links?

Starting from 2019 R3 downloader config files located in model's folder

And where it located? How am I able to use downloader without downloading whole openvino. In RC2 and below I could do that.

[vladimir-dudnik]

That's how it work for me:

1. I've cloned open_model_zoo repo from github by 2019_R3 tag
2. cd to open_model_zoo/tools/downloader folder
3. run command python3 downloader.py --print_all from there, and it works

[banderlog]

It works:

git clone -b '2019_R3' --single-branch --depth 1 https://github.com/opencv/open_model_zoo
cd open_model_zoo/tools/downloader
./downloader.py --print_all
>>action-recognition-0001-decoder
>>action-recognition-0001-encoder
>>age-gender-recognition-retail-0013
>>...

And to answer my question:

https://github.com/opencv/open_model_zoo/blob/acf297c73db8cb3f68791ae1fad4a7cc4a6039e5/tools/downloader/downloader.py#L212

https://github.com/opencv/open_model_zoo/blob/acf297c73db8cb3f68791ae1fad4a7cc4a6039e5/tools/downloader/common.py#L390-L398

Basically, it parses subdirs in ../../models/* relative to downloader.py

So, before you was able to download downloader separetely from model_zoo, now not.