Yesterday, a developer published a card on technical community V2EX, told about the is between tears and smiles experience when oneself are studying deepness study to everybody.
This developer expresses, after be installed in computer and running Fastai (the implementation library that Fastai is machine study and deepness study algorithm not only, still one is cover to it corresponding course and tutorial) , no matter the model of place training has many simple, can bring about a system appear dead machine and restarted circumstance. This developer returns respecting, run alone with Tensorflow/pytorch simple case won't have a problem however.
Because the be used to of developer has thinking, after party discovers a problem also Debug a long time, but all relevant environments were checked in the platoon (after CUDA/cuDNN/Anaconda/PyTorch) , did not discover the germ of the problem is in. But under, this developer goes the GitHub storehouse of Fastai carried an Issue. Because,discovering this problem finally is of their company purchase spelling what inferior power source buys to cause on great platform.
The respecting in the return that the developer that this ID is Geekan leaves in Issue before this (at present content already by him developer re-edit) , restarted problem is from go all out great on bought 550W power supply brings about. This developer expresses, nominal the power source of power 550W, only actually 200W is controlled, suspicion was to buy fake. He changed finally power source, state the problem has been solved.
Nevertheless we also are in the reply of this Issue the sound that saw a few doubt, think this is black go all out great " Hei Wen " . To this, you are how of look upon, the welcome leaves your think of a way in the comment.
