core dumped

[xxx2974]

Hello！when I run the 'train_market1501.sh' file,the training process can be done,but core dumped occurred during the testing process. How can I solve the problem?
I use two gpus. The details are:
......
Epoch: [750][100/100] Time 1.331 (2.397) Data 0.000 (0.016) Loss 0.000 (0.000) Prec 100.00% (100.00%) nEff 0.00 (0.02) nA) effRatio 0.00% (0.00%)

• Finished epoch 750

Test with model epoch_750:
=> Loaded checkpoint 'logs/market1501/tmp/epoch_750.pth.tar'
/part_bilinear_reid/reid/feature_extraction/cnn.py:12: UserWarning: volatile was removed and now has no effect. Use `with t
inputs = Variable(inputs, volatile=True)
Extract Features: [10/78] Time 1.644 (3.047) Data 0.000 (0.463)
Extract Features: [20/78] Time 1.160 (2.375) Data 0.000 (0.232)
Extract Features: [30/78] Time 1.552 (2.279) Data 0.000 (0.155)
Extract Features: [40/78] Time 1.153 (2.158) Data 0.000 (0.116)
Extract Features: [50/78] Time 1.273 (2.038) Data 0.000 (0.093)
Extract Features: [60/78] Time 1.129 (1.974) Data 0.001 (0.078)
Extract Features: [70/78] Time 1.094 (1.904) Data 0.001 (0.067)
./run.sh: 行 46: 15770 段错误 CUDA_VISIBLE_DEVICES=$GPU_ID $PYTHON train.py $STR_PARAM
@yuminsuh

[yuminsuh]

Hi, it is hard to know where the error occurred from this log.. Could you tell me in which line of reid/evaluators.py the error occurred?

[xxx2974]

Thanks for your apply. I use simple print statement and I find that the line of"dist.addmm_(1, -2, x, y.t())" may have an error.
for example, "print("9-Where is the core dumped?")" can be printed,but " print("10-Where is the core dumped?")" can't be .
print("9-Where is the core dumped?")
dist.addmm_(1, -2, x, y.t())
print("10-Where is the core dumped?")

[xxx2974]

@yuminsuh
Now the core dumped problem has been solved. I installed the pytorch version 1.0.0 instead of 0.4. It can work,and I didn't know the specific reason.