can only detect "smoke"

[willswordh]

Hi great work!
I used to your model to process video, but it can only output bounding box of smoke. Not sure I am doing it right, could you help me? Thanks a lot!!!

I used labels:
labels = {1: 'fire', 2: 'smoke'}

[X-Men01]

Hi @willswordh

Thank you for reaching out, and I'm happy to assist!

It seems the issue might be related to the label mapping. Based on the model configuration, the correct label mapping is:

labels = {0: 'Fire', 1: 'Smoke'}

Please update your code to use this mapping instead of {1: 'fire', 2: 'smoke'}. This will ensure the class IDs from the model output are matched properly to their respective labels.

Let me know if you need further assistance!

Best regards,