Konstantinos Monachopoulos

It is not about the images, it is about the error metric. I suppose I have to make a script to inference through the VOC.

I am using the classification script, with the correct configuration parameters on inceptionv3 and v4 and get really bad results. Even single image inference gives wrong results (miss-classified). The same...

So, what .txt file did you use for the annotation ?? This is a reference of the file I am using : ``` 1: 'goldfish, Carassius auratus', 2: 'great white...

I still get wrong results with that annotation list .. I have tried lot of annotation lists that i found online (including the one that you gave me) and it...

Hello @MichaelDylan77, In the inference script you will see that the output of the network before passing the information to the decoder is : `raw_output_up = tf.argmax(raw_output_up, axis=3)` Converting the...

Pass them through a sigmoid function .. maybe that will help!

I think this is because this is not supported in the code yet. Because the code takes only one value into consideration which is the class probability with the maximum...

Try replacing the function `process_predicts` with the function below in `demo.py` : ``` # Interpret network output (7x7x30) to bb coordinates (long process) def InterpretPredictions(predicts): p_classes = predicts[0, :, :,...

If you want the network to predict only one class (true / false), you should re-train it for that class only. I you want to use it as it is...

Same problem when using a batch > 1 ... > Invalid indices: [43200,0] = [1, 90, 0, 0] does not index into [10,90,120,64]