This is the first time I am launching the CIFAR-10 classification from Tensorflow on my laptop with the GeForce 940M. I start a workout with predefined parameters as follows:
python cifar10_train.py
after step 1800 I get the following errors:
E tensorflow/stream_executor/cuda/cuda_event.cc:33] Error polling for event status: failed to query event: CUDA_ERROR_ILLEGAL_ADDRESS
E tensorflow/stream_executor/cuda/cuda_driver.cc:1182] failed to enqueue async memcpy from device to host: CUDA_ERROR_ILLEGAL_ADDRESS; host dst: 0x7ff8e9bf26c0; GPU src: 0x5011c0600; size: 16=0x10
F tensorflow/core/common_runtime/gpu/gpu_event_mgr.cc:105] Unexpected Event status: 1
I tensorflow/stream_executor/stream.cc:3304] stream 0x35e7190 did not block host until done; was already in an error state
Aborted (core dumped)
Does anyone have any ideas?
Many thanks for your help! Any advice is kindly appreciated!
source
share