hxuaj / tf2-faster-rcnn Goto Github PK

Hi at first I want to say that is the best Faster RCNN I could find which is created with Tensorflow2.

I try this one and change some parts. Special the classes and the database. But nothing more. In my case I take a look to the output and all the data is labled with the background index. You can imagine, the result are no labels. Do you have any idea why this should be happen? The interface to fill the FasterRCNN class is similar to yours. The label count in each of the images is around 10 to 20 labels, the size of the images are 1000:840.
Edit: it looks like the problem of the RPN which is decribed in https://arxiv.org/pdf/1506.01497.pdf 3.1.3. Is there any algorithm implemented?

File "train.py", line 140, in
ap ,mAP = voc_eval(eval_ds, model, ds.num_classes)
File "C:\Users\anind\Documents\Dissertation Codes\PhaseII_Novelty\tf2-faster-rcnn\eval.py", line 205, in voc_eval
precision, recall = voc_eval_pre_rec(dataset, model, num_classes)
File "C:\Users\anind\Documents\Dissertation Codes\PhaseII_Novelty\tf2-faster-rcnn\eval.py", line 35, in voc_eval_pre_rec
labels, bboxes, cls_prob = model(img_input)
ValueError: too many values to unpack (expected 3)

#1 Regarding this issue, my current workaround is to create a new model with is_training=False and copy weights into the new model. Not sure if I did something wrong, since I am using TensorFlow r2.11.0.

temp_model = FasterRCNN(is_training=False)
temp_model(init_model())
temp_model.set_weights(model.get_weights())
ap ,mAP = voc_eval(eval_ds, temp_model, ds.num_classes)

Can I use the same model to train a new dataset with only 2 classes ?
if yes, what shall I take into consideration to change the model to fit for the new data ?