Stefan

I've adapted the generated PHP-Code, so that it works with 5.3. It was quite easy. Thanks for looking into the issue. You're right. It might be worth a thought to...

I have the same issue, when I use the input_tensor(!)-Parameter: ``` import efficientnet.keras as efn base_model = efn.EfficientNetB0( include_top=False, weights=None, input_tensor=Input(shape=(244, 244, 3))) ``` By the way, I'm using tensorflow...

Hi NFiedler, I just stumbled over your comment and desperatly need to zoom some large images in ImageTagger. Did you already tried to integrate EasyZoom into ImageTagger or do you...

Thanks @NFiedler for your quick answer and the hint to annotation.js I will look into it. Sounds like a good idea to reimplement the frontend in a proper JS framework....

I would be interested in this as well. Were you able to retrain your model? Thanks for providing this repo.

Last project I used the pixel annotation tool: https://github.com/abreheret/PixelAnnotationTool You really have to get used to it, but it works.

Just have a look at train_mvod_basenet.py. You also need to provide the path to a cache-directory when you run train_mvod_basenet.py with `--cache_path /tmp`

You are right. That was a python3-only solution ``` #if PY_MAJOR_VERSION >= 3 static struct PyModuleDef cModPyYolo = { PyModuleDef_HEAD_INIT, "pyyolo", "", -1, PyyoloMethods }; PyMODINIT_FUNC PyInit_pyyolo(void) { return PyModule_Create(&cModPyYolo);...

Just to mention: the file cifar10.npz is created in the "root"-folder, not in the data-subdirectory (where it is expected by cae.py). So you have to move it to the data-directory....

Just from memories this has run in around 20 minutes on nvidia 1070. Without gpu ... 6 hours?