Issue with Assertion Error in Label Generation

[honeius]

Dear developers,

First, thank you for your hard work and for making this project available. I greatly appreciate your contributions.

I am currently working with your CcGAN implementation for a custom dataset . While the example provided works smoothly, I encountered an issue when applying my dataset. Specifically, I am facing an AssertionError during the label generation process (run_gene.sh). Here is a brief overview of the situation:

1. Problem Description:
   The error occurs in the label generation part, where the following assertion fails:

---

Traceback (most recent call last):
  File "/path/to/main.py", line 780, in <module>
    assert fake_labels.max()>1
AssertionError

---

I reviewed the some functions to ensure the labels are generated correctly, but I haven't found any clear reason why the labels are failing the condition.

2. Dataset and Label Information:

The labels range from:

---

1963 images left and there are 80 unique labels

 Range of unnormalized labels: (0.009999999776482582,0.8100000023841858)
 Range of normalized labels: (0.02439024329300724,1.0)
 There are 80 unique labels.

---

The label generation logic seems to apply correctly based on the raw dataset, but the normalized or generated fake_labels do not meet the condition.

Would you be able to provide some guidance on why the fake_labels might not be reaching the expected range, and how I can resolve this AssertionError?

Thank you for your time and help!