Kernel dies when trying to run tiny-yolo-image-loop

[xilinx]

Kernel dies when trying to run tiny-yolo-image-loop

[giuliogamba]

Hi,
Actually we never experienced that, but I assume it can be due to the device going out-of-memory.
In general, you can check why a notebook kernel dies ny checking the log under /var/log/jupyter.log

[MinLiAmoy]

same problem:D

[shwetad7]

Hey giuliogamba,
Looks like some memory allocation/deallocation issue. Following is content copied from jupyter.log for your reference.

[I 00:21:18.864 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[I 00:21:18.938 NotebookApp] 302 GET / (127.0.0.1) 4.79ms
[I 00:21:18.950 NotebookApp] 302 GET /tree? (127.0.0.1) 8.19ms
[W 00:21:34.353 NotebookApp] 404 GET /api/kernels/9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c/channels?session_id=3b9bb37d5bff4e0fbb0290deca8c577c (192.168.40.193): Kernel does not exist: 9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c
[W 00:21:34.445 NotebookApp] 404 GET /api/kernels/9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c/channels?session_id=3b9bb37d5bff4e0fbb0290deca8c577c (192.168.40.193) 213.01ms referer=None
[W 00:21:37.617 NotebookApp] Replacing stale connection: 9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c:3b9bb37d5bff4e0fbb0290deca8c577c
[W 00:22:06.471 NotebookApp] Replacing stale connection: 9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c:3b9bb37d5bff4e0fbb0290deca8c577c
[W 00:23:30.480 NotebookApp] Replacing stale connection: 9b30aaf4-0ba0-471d-a2f9-59ffe27fda6c:3b9bb37d5bff4e0fbb0290deca8c577c
[I 00:26:09.685 NotebookApp] Kernel started: 5e9785a8-c5b5-452f-aa6a-5c8ce6b80f21
[I 00:26:18.436 NotebookApp] Adapting to protocol v5.1 for kernel 5e9785a8-c5b5-452f-aa6a-5c8ce6b80f21
layer filters size input output
0 conv 16 3 x 3 / 2 416 x 416 x 3 -> 208 x 208 x 16
1 conv 64 3 x 3 / 1 208 x 208 x 16 -> 208 x 208 x 64
2 max 2 x 2 / 2 208 x 208 x 64 -> 104 x 104 x 64
3 conv 64 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 64
4 max 2 x 2 / 2 104 x 104 x 64 -> 52 x 52 x 64
5 conv 128 3 x 3 / 1 52 x 52 x 64 -> 52 x 52 x 128
6 max 2 x 2 / 2 52 x 52 x 128 -> 26 x 26 x 128
7 conv 256 3 x 3 / 1 26 x 26 x 128 -> 26 x 26 x 256
8 max 2 x 2 / 2 26 x 26 x 256 -> 13 x 13 x 256
9 conv 512 3 x 3 / 1 13 x 13 x 256 -> 13 x 13 x 512
10 max 2 x 2 / 1 13 x 13 x 512 -> 13 x 13 x 512
11 conv 512 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x 512
12 conv 512 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x 512
13 conv 125 1 x 1 / 1 13 x 13 x 512 -> 13 x 13 x 125
14 detection
mask_scale: Using default '1.000000'
[I 00:30:09.582 NotebookApp] Saving file at /qnn/HDMI_Yolo-withcam.ipynb
[I 00:32:09.594 NotebookApp] Saving file at /qnn/HDMI_Yolo-withcam.ipynb
layer filters size input output
0 conv 16 3 x 3 / 2 416 x 416 x 3 -> 208 x 208 x 16
1 conv 64 3 x 3 / 1 208 x 208 x 16 -> 208 x 208 x 64
2 max 2 x 2 / 2 208 x 208 x 64 -> 104 x 104 x 64
3 conv 64 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 64
4 max 2 x 2 / 2 104 x 104 x 64 -> 52 x 52 x 64
5 conv 128 3 x 3 / 1 52 x 52 x 64 -> 52 x 52 x 128
6 max 2 x 2 / 2 52 x 52 x 128 -> 26 x 26 x 128
7 conv 256 3 x 3 / 1 26 x 26 x 128 -> 26 x 26 x 256
8 max 2 x 2 / 2 26 x 26 x 256 -> 13 x 13 x 256
9 conv 512 3 x 3 / 1 13 x 13 x 256 -> 13 x 13 x 512
10 max 2 x 2 / 1 13 x 13 x 512 -> 13 x 13 x 512
11 conv 512 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x 512
12 conv 512 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x 512
13 conv 125 1 x 1 / 1 13 x 13 x 512 -> 13 x 13 x 125
14 detection
mask_scale: Using default '1.000000'
*** Error in `/opt/python3.6/bin/python3.6': double free or corruption (out): 0x192fe008 ***
[I 00:34:54.661 NotebookApp] KernelRestarter: restarting kernel (1/5), keep random ports
kernel 5e9785a8-c5b5-452f-aa6a-5c8ce6b80f21 restarted

[giuliogamba]

Hi,

Is this happening in the original notebook?
It seems you are calling the initialization of darknet lib.parse_network_cfg (performing memory allocation for the layers) twice. This might end-up in double memory allocation.

[shwetad7]

But I have called lib.parse_network_cfg only once. Is it called internally in any other function ?
Is there a way to free this memory and allocate again?

[giuliogamba]

You might want to check the call free_network in the darknet lib.