Video2X in Colab crashing with T4 GPU #635
Unanswered
tinyhands01
asked this question in
Q&A
Replies: 1 comment 3 replies
-
Theoretically it shouldn't crash no matter which GPU you use. I've also seen other reports like #630 reporting similar behaviors. RAM usage looks normal, GPU usage looks normal, storage shouldn't be a problem, so I've not narrowed the problem down either. |
Beta Was this translation helpful? Give feedback.
3 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'll preface by saying that I know next to nothing about Video2X and even less about Google Colab, but I can generally follow instructions pretty well and appear to have managed to make it work.
At least until I got greedy and started poking around at what other GPU options were available for free in Colab and it turns out that if you do a 'Factory Reset Runtime' a couple of times you might randomly get a better GPU than the stock K80. So I did that and got a T4 processor and, wanting to see the difference, rebuilt the dependencies and installed.
The T4 was getting 4.45 FPS, so a significant improvement over the K80's 1.13, except that the T4 would consistently crash on or before the 119th frame. The K80 hasn't crashed (yet - 1300 frames & counting) so I don't want to point the finger at my source file.
Does anyone know why the T4 keeps crashing? Is there something in the code that needs to change to make it work on the T4 architecture? Before you ask me for logs, please list detailed steps to generate exactly what you want to see. Please also see if you can replicate the crash using your own source file and the T4 processor.
TIA
Beta Was this translation helpful? Give feedback.
All reactions