Issues while running Zero Shot Prediction

[alexgcsa]

Hi,

Thanks for the hard work on making this available for everyone :)

I am having two issues while running the zero shot predictor (for predicting ejection fraction):

• Every time I run the Python script, I get a different answer, even for the example. I have just run it 5 times, and the outputs were:

1. Predicted ejection fraction is 19.8%
2. Predicted ejection fraction is 19.6%
3. Predicted ejection fraction is 20.0%
4. Predicted ejection fraction is 19.0%
5. Predicted ejection fraction is 20.5%

Why is this happening? Shouldn't it be deterministic?

• I am trying to replace the example file with other files. However, it is not always that I get a results and I am experiencing the following error:

OpenCV(4.11.0) /io/opencv/modules/videoio/src/cap_images.cpp:267: error: (-215:Assertion failed) number < max_number in function 'icvExtractPattern'


Traceback (most recent call last):
  File "/home/alexgcsa/Downloads/echo_CLIP-main/zero_shot_example.py", line 29, in <module>
    test_video = torch.stack(
                 ^^^^^^^^^^^^
RuntimeError: stack expects a non-empty TensorList

Do you have any recommendations on how to fix this issue?