mmr/tkDNN
Archived
1
0
Fork 0
Code Issues 46 Pull Requests 12 Releases 4 Activity

Fails while exporting yolov4 #3

New Issue
Closed
opened 2020-04-29 15:45:48 +02:00 by Kmarconi · 3 comments
Kmarconi commented 2020-04-29 15:45:48 +02:00 (Migrated from github.com)
Copy Link
Copy Source

Hi ! FIrst thanks for your repo, I saw what you've done with the new yolov4 weights in terms of FPS and I wanted to reproduce it. I'm working on a Jetson Xavier and when I'm trying to export the yolov4 weight, I got this error at for the 108 layer : darknet: ./src/darknet.c:519: run_export: Assertion `0 == "layer type not supported for export"' failed. Do you know why this error is showing up ?

Full error :
n: 108, type 3 darknet: ./src/darknet.c:519: run_export: Assertion 0 == "layer type not supported for export"' failed. Aborted

Hi ! FIrst thanks for your repo, I saw what you've done with the new yolov4 weights in terms of FPS and I wanted to reproduce it. I'm working on a Jetson Xavier and when I'm trying to export the yolov4 weight, I got this error at for the 108 layer : darknet: ./src/darknet.c:519: run_export: Assertion `0 == "layer type not supported for export"' failed. Do you know why this error is showing up ? Full error : `n: 108, type 3 darknet: ./src/darknet.c:519: run_export: Assertion 0 == "layer type not supported for export"' failed. Aborted`
Kmarconi commented 2020-04-29 16:22:19 +02:00 (Migrated from github.com)
Copy Link
Copy Source

By commenting out the 519 line of the darknet.c file, i'm able to export my weights, don't know if t's a good idea

By commenting out the 519 line of the darknet.c file, i'm able to export my weights, don't know if t's a good idea
ceccocats commented 2020-04-29 17:57:22 +02:00 (Migrated from github.com)
Copy Link
Copy Source

Hi,
The export in our darknet fork was wrong (it missed a layer in the export function).
We fixed it and now should work.
Please check you are using the correct repo:
https://git.hipert.unimore.it/fgatti/darknet

Anyway if you need to test yolo4 with the official weights you dont need to export.
You just run the ./test_yolo4 sample and it will automatically download it.

Hi, The export in our darknet fork was wrong (it missed a layer in the export function). We fixed it and now should work. Please check you are using the correct repo: https://git.hipert.unimore.it/fgatti/darknet Anyway if you need to test yolo4 with the official weights you dont need to export. You just run the ./test_yolo4 sample and it will automatically download it.
Kmarconi commented 2020-04-30 09:11:22 +02:00 (Migrated from github.com)
Copy Link
Copy Source

Thanks for your reply ! :) Yeah i saw yesterday that by doing directly ./test_yolo4 it is autommaticaly downloading it :) Just so you know, I'm arround 34 FPS on my Xavier with yolov4

Thanks for your reply ! :) Yeah i saw yesterday that by doing directly ./test_yolo4 it is autommaticaly downloading it :) Just so you know, I'm arround 34 FPS on my Xavier with yolov4
This repo is archived. You cannot comment on issues.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
bug
duplicate
enhancement
explanation
help wanted
invalid
question
wontfix
No Label
Milestone
No items
No Milestone
Assignees
Clear assignees
eeeck (Erick)
No Assignees
1 Participants
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: mmr/tkDNN#3
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?