Open Jorisfournel opened 7 years ago
I get the following error :
VNet.py:53: RuntimeWarning: divide by zero encountered in divide weightData[defLab == 1] = np.prod(defLab.shape) / np.sum((defLab==1).astype(dtype=np.float32)) F1018 14:52:56.896700 11971 syncedmem.cpp:51] Check failed: error == cudaSuccess (2 vs. 0) out of memory Check failure stack trace: Abandon (core dumped)
Any idea ? Can the divide by zero be linked to the out of memory error or is it just that my GPU memory is too small ?
I have the same error as yours. And have you solved your problem? If solved, I'll appreciate it when you can share your solution with me. Thank you!
Try adding smoothing parameter in denominator
I get the following error :
VNet.py:53: RuntimeWarning: divide by zero encountered in divide weightData[defLab == 1] = np.prod(defLab.shape) / np.sum((defLab==1).astype(dtype=np.float32)) F1018 14:52:56.896700 11971 syncedmem.cpp:51] Check failed: error == cudaSuccess (2 vs. 0) out of memory Check failure stack trace: Abandon (core dumped)
Any idea ? Can the divide by zero be linked to the out of memory error or is it just that my GPU memory is too small ?