Deepak
Did you able to fix this ?
@PurvangL What is the command that you gave in both the machines? I was getting similar error due to wrong values that I gave in `num_machines` and `machine_rank`
AFAIK, there is no restriction on maximal object detection. In my finetuned model I am getting detection's upto 600 objects from single frame.
@Lihi-Gur-Arie For Evaluation: I think you can modify the threshold in the pycocotools py files. In this file `pycocotools/cocoeval.py`, you can update 100 to 300. For training: You can modify...
@Lihi-Gur-Arie I am unsure about the error, as I have not modified this `maxDets` param for my use. But I guess, If you modified only in line 508, it will...
Not used any extra augmentation apart from what is available in train.py
@Moji0110 @ahmettkilc You can download crowdhuman dataset from here: https://www.crowdhuman.org/download.html
Please refer detect.py
May be its due to you are using binary_crossentropy, if you are using categorical_crossentropy this may be resolved.