Hello, thank you for sharing your code. I encountered the following problem during training suddenly:
[W CUDAGuardImpl.h:124] Warning: CUDA warning: unknown error (function destroyEvent)
Traceback (most recent cal last):
File "scripts/segmentation_train.py", line 89, in
main()
File "scripts/segmentation_train.py", line 45, in main
TrainLoop(
File "E:\zll\Diffusion-based-Segmentation-mainl.\guided_diffusion\train_util.py", line 186, in run_loop
self .run_step(batch, cond)
File "E:\zll\Diffusion-based-Segmentation-main.\guided_diffusion\train_util.py", line 207, in run_step
sample = self.forward_backward(batch, cond)
File "E:\zll\Diffusionbased-Segmentation-main.\guided_diffusion\train_util.py", line 256, in forward_backward
self.mp_trainer.backward(loss)
File "E:\zll\Diffusion-based-Segmentation-main.\guided_diffusion\fp16_util.py", line 181, in backward
My computer's CUDA version is 12.2, while the CUDA version in the virtual environment is 11.7, and PyTorch version is 2.0.1.
Hello, thank you for sharing your code. I encountered the following problem during training suddenly: [W CUDAGuardImpl.h:124] Warning: CUDA warning: unknown error (function destroyEvent) Traceback (most recent cal last): File "scripts/segmentation_train.py", line 89, in
main()
File "scripts/segmentation_train.py", line 45, in main
TrainLoop(
File "E:\zll\Diffusion-based-Segmentation-mainl.\guided_diffusion\train_util.py", line 186, in run_loop
self .run_step(batch, cond)
File "E:\zll\Diffusion-based-Segmentation-main.\guided_diffusion\train_util.py", line 207, in run_step
sample = self.forward_backward(batch, cond)
File "E:\zll\Diffusionbased-Segmentation-main.\guided_diffusion\train_util.py", line 256, in forward_backward
self.mp_trainer.backward(loss)
File "E:\zll\Diffusion-based-Segmentation-main.\guided_diffusion\fp16_util.py", line 181, in backward
My computer's CUDA version is 12.2, while the CUDA version in the virtual environment is 11.7, and PyTorch version is 2.0.1.