Hey @XinzeLee, I followed Rotate-tutorial 1 colab to train the model for 100 epochs but I observe the same low number of labels predicted at each epoch:
I have also tried it for my custom dataset & a similar problem persists. What are your thoughts on it?
Hey @XinzeLee, I followed Rotate-tutorial 1 colab to train the model for 100 epochs but I observe the same low number of labels predicted at each epoch:
I have also tried it for my custom dataset & a similar problem persists. What are your thoughts on it?