[MvHEAT-DET] Problems encountered during training and testing

[Peipiggy]

Thanks a lot for your great dataset and work! I have encountered the following issues during reproduction:

1. Training issue: When I tried to train the model on my device, I ran into CUDA OOM errors. Could you please provide the details about the hardware you used for training, especially the memory capacity?
2. Testing issue: When I tried to test the model with the pre-trained ckpt you provided (https://github.com/Event-AHU/OpenEvDET/tree/main/MvHeatDET#checkpoint-download) , I encountered the error "RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory". According to the solutions on web, the error happens when the ckpt file is corrupted and needs to be re-downloaded. However, I have tried downloading from both Baiduyun and Dropbox, and the error persists. Could you please help me check what's the matter with the file or code?

Thanks again for your great work and kind help!