Hello! I encountered an error during training, but the program runs normally when reproducing results. The error message is as follows:
复制
RuntimeError: nvrtc: error: invalid value for --gpu-architecture (-arch)
nvrtc compilation failed:
#define NAN __int_as_float(0x7fffffff)
#define POS_INFINITY __int_as_float(0x7f800000)
#define NEG_INFINITY __int_as_float(0xff800000)
My running environment is CUDA 11.3 and PyTorch 1.10.
Hello! I encountered an error during training, but the program runs normally when reproducing results. The error message is as follows:
复制
RuntimeError: nvrtc: error: invalid value for --gpu-architecture (-arch)
nvrtc compilation failed:
#define NAN __int_as_float(0x7fffffff)
#define POS_INFINITY __int_as_float(0x7f800000)
#define NEG_INFINITY __int_as_float(0xff800000)
My running environment is CUDA 11.3 and PyTorch 1.10.