DLS/Course 4/Week 3/reboot doesn't help for problem CUDA runtime implicit

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hello everyone,

I’m working on the second homework of course 4 week 3, and I also meet the problem of

CUDA runtime implicit initialization on GPU:0 failed. Status: out of memory

The problem occur when running the code you have provided, on the 4 th code block of doing this:
image_list_ds = tf.data.Dataset.list_files(image_list, shuffle=False)

I have already tried adding
tf.keras.backend.clear_session(),
restart the kernel ,
help button to reboot the server
and none of these helps.

So how could solve this problem and continue to do the homework?

Best Wishes,

Sec

1 Like
2

I’m encountering the same problem tried all these solutions aswell were you able to find a solution that works?

1 Like
3

I don’t have any troubles running the code in that cell. Are you running the notebook in Coursera’s environment through your browser? Or are you running the notebook locally?

Do not add any code.

It should be sufficient to use “Kernel → Restart & Clear Output”, followed by “Cell → Run All”.

4

Same problem for me, and your solution does not work.

5

Same here. I am glad I am not alone on this, has been killing me all morning/afternoon

6

I did not claim to have a solution. I suggested that adding code to the notebook and restarting the server are not expected to be helpful.

I’ll ask the course staff to investigate.

7

@Sectum Based on this thread, what did coursera help tell you?

8

I also had the same problem. A reboot of the server solved it for me. (“Help” in the right top corner which opens the Lab Help => Reboot server.)