Darknet: Couldn't open file:/backup//yolo-obj_100.weights

[vvkv]

I am training my yolo network on an ubuntu machine, and the training seems to go well for the first 100 steps after which it gets aborted with the error: "Couldn't open file:/backup//yolo-obj_100.weights"

Here is the terminal output for reference:

ubuntu@ip-172-31-40-216:~/darknet$ ./darknet detector train cfg/obj.names cfg/yolo-obj.cfg darknet19_448.conv.23
train: Using default 'data/train.list'
backup: Using default '/backup/'
yolo-obj
layer filters size input output
0 conv 32 3 x 3 / 1 416 x 416 x 3 -> 416 x 416 x 32
1 max 2 x 2 / 2 416 x 416 x 32 -> 208 x 208 x 32
2 conv 64 3 x 3 / 1 208 x 208 x 32 -> 208 x 208 x 64
3 max 2 x 2 / 2 208 x 208 x 64 -> 104 x 104 x 64
4 conv 128 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 128
5 conv 64 1 x 1 / 1 104 x 104 x 128 -> 104 x 104 x 64
6 conv 128 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 128
7 max 2 x 2 / 2 104 x 104 x 128 -> 52 x 52 x 128
8 conv 256 3 x 3 / 1 52 x 52 x 128 -> 52 x 52 x 256
9 conv 128 1 x 1 / 1 52 x 52 x 256 -> 52 x 52 x 128
10 conv 256 3 x 3 / 1 52 x 52 x 128 -> 52 x 52 x 256
11 max 2 x 2 / 2 52 x 52 x 256 -> 26 x 26 x 256
12 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
13 conv 256 1 x 1 / 1 26 x 26 x 512 -> 26 x 26 x 256
14 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
15 conv 256 1 x 1 / 1 26 x 26 x 512 -> 26 x 26 x 256
16 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
17 max 2 x 2 / 2 26 x 26 x 512 -> 13 x 13 x 512
18 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
19 conv 512 1 x 1 / 1 13 x 13 x1024 -> 13 x 13 x 512
20 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
21 conv 512 1 x 1 / 1 13 x 13 x1024 -> 13 x 13 x 512
22 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
23 conv 1024 3 x 3 / 1 13 x 13 x1024 -> 13 x 13 x1024
24 conv 1024 3 x 3 / 1 13 x 13 x1024 -> 13 x 13 x1024
25 route 16
26 reorg / 2 26 x 26 x 512 -> 13 x 13 x2048
27 route 26 24
28 conv 35 3 x 3 / 1 13 x 13 x3072 -> 13 x 13 x 35
29 conv 35 1 x 1 / 1 13 x 13 x 35 -> 13 x 13 x 35
30 detection
Loading weights from darknet19_448.conv.23...
seen 32
Done!
Learning Rate: 0.0001, Momentum: 0.9, Decay: 0.0005
Loaded: 1.204687 seconds
Region Avg IOU: 0.327445, Class: 0.507517, Obj: 0.602026, No Obj: 0.534384, Avg Recall: 0.125000, count: 8
Region Avg IOU: 0.373777, Class: 0.478211, Obj: 0.561162, No Obj: 0.535185, Avg Recall: 0.375000, count: 8
Region Avg IOU: 0.214813, Class: 0.614562, Obj: 0.524953, No Obj: 0.534205, Avg Recall: 0.000000, count: 8
Region Avg IOU: 0.314540, Class: 0.647449, Obj: 0.504414, No Obj: 0.535127, Avg Recall: 0.250000, count: 8
Region Avg IOU: 0.286844, Class: 0.568035, Obj: 0.508885, No Obj: 0.533469, Avg Recall: 0.250000, count: 8
Region Avg IOU: 0.214517, Class: 0.626371, Obj: 0.443656, No Obj: 0.534725, Avg Recall: 0.000000, count: 8
.
.
.
.
Region Avg IOU: 0.660878, Class: 0.873684, Obj: 0.239481, No Obj: 0.060791, Avg Recall: 1.000000, count: 8
Region Avg IOU: 0.702378, Class: 0.806532, Obj: 0.313847, No Obj: 0.060674, Avg Recall: 1.000000, count: 8
100: 0.368687, 0.404861 avg, 0.001000 rate, 7.345735 seconds, 6400 images
Saving weights to /backup//yolo-obj_100.weights
Couldn't open file: /backup//yolo-obj_100.weights

Why is this happening? I have the backup folder created at the same directory level as the darknet exe, so, darknet/backup, and the folder is empty.

Any guidance will be very useful

[CBIR-LL]

Manully create the 'backup' folder and retrain.

[tgwagner]

I ran into the same problem. go to your /cfg/yolo.data file and make sure you have a line that says backup = backup

(not backup = backup/) the code will append "/theweightsfile.weights" to this directory and you'll have double slashes.

Good luck.

[michaelfuchs1]

just had the same problem.

although:

• my "backup" folder existed
• the line in /cfg/yolo.data was "backup = backup"

turned out it was a stupid mistake on my side: I was executing the command from my "C:Usersmicha" directory, instead of the correct "C:usersmicha....darknet-master".

HTH,
Michael

[GioLomia]

I have the same issue. Where you able to solve it?

[lsd1994]

I have the same issue. Where you able to solve it?

Just remove the backup folder and create a new one

[GioLomia]

I have the same issue. Where you able to solve it?

Just remove the backup folder and create a new one

I did but it still did not work

[hemanthkumar3111]

having the same issue, even after recreating the backup folder. any tips are really appreciated

[GLorieul]

The problem is the leading / which means that the path refers to a directory named backup and located at the root of the machine. Which most certainly does not exist on your machine…

Unlike some have said, the problem is not the double / in /backup//yolo-obj_100.weights: at least in Linux, // is equivalent to /./. Actually I like setting backup = ./backup/ in the configuration file because I find it looks good and unambiguous, which in my case eventually expands as Saving weights to ./backup//yolo-obj_last.weights… And it works!

There can be three things that you are doing wrong:

1. Your object data configuration file specifies a path that does not correspond to an actual directory.
2. That directory exists, but you don't have the required permissions e.g. read-only directory.
3. Your object data configuration file uses a key that is misspelled (e.g. bakup instead of backup). If so, darknet will use its default value: /backup/. (This was my case)

How to recognize whether you are in case #1, #2 or #3: If you are in case #3, darknet will display backup: Using default '/backup/' in one of the first lines of its prompt (in particular, before loading the network's configuration).

In case #1 or case #2: do as @CBIR-LL proposed: (i) delete existing directory and create a fresh one, then (ii) double check that the path you specified is indeed correct.

In case #3: fix your configuration file by changing the key name to an actual legal name e.g. backup instead of bakup. Assure yourself that you did it right by checking the absence of the backup: Using default '/backup/' line on your next run of darknet.

[USNA2014]

The reason is because remember the darknet.exe program being run on cmd prompt needs all paths relative to it. For example, my backup folder was actually in the x64 folder. So I changed it to to x64/backup/ (the final slash is ok since it's opening that folder).

[dippa-1]

I ran into the same problem. go to your /cfg/yolo.data file and make sure you have a line that says backup = backup

(not backup = backup/) the code will append "/theweightsfile.weights" to this directory and you'll have double slashes.

Good luck.

worked for me, thanks

[shouryasimha]

@GLorieul Hi, I'm trying to run my model on google colab and gave by a backup path to a folder that I created in google colab but the model stops abruptly at 100th iteration.

Please help me out to solve this issue as I have been stuck with this issue for a week.

[PayamBeirami]

I ran into the same problem. go to your /cfg/yolo.data file and make sure you have a line that says backup = backup

(not backup = backup/) the code will append "/theweightsfile.weights" to this directory and you'll have double slashes.

Good luck.

i got the same issue and was about to try your advise but i see theres no yolo.data file in cfg folder .

[USNA2014]

@shouryasimha hello, can you say where your backup folder is relative to your darknet folder? This seems to be an issue of the program not able to find the backup folder. Sometimes darknet.exe ends abruptly without displaying an error.

[shouryasimha]

It was solved by using sudo doc2unix. Can you help me to find a way to
visualise the training results?

On Mon, Apr 6, 2020 at 1:12 PM GS [e-mail] wrote:

@shouryasimha https://github.com/shouryasimha hello, can you say where
your backup folder is relative to your darknet folder? This seems to be an
issue of the program not able to find the backup folder. Sometimes
darknet.exe ends abruptly without displaying an error.

—
You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub
https://github.com/pjreddie/darknet/issues/416#issuecomment-609922420,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2FYMVGZZYYIOEAYSMGDRLIEPDANCNFSM4EMBKEOQ
.

[USNA2014]

@shouryasimha also why is it yolov3.backup? Backup shouldn’t be a file extension and should be a folder. Like this yolov3/backup

[shouryasimha]

Yolov3.backup is the weight corresponding to the latest weights wrt to the
epochs , such as if your model stopped at 7123 iteration, using that
yolov3.backup you can resume training from there . It’s helpful if you are
running your model on colab or so , in case runtime gets disconnected

On Mon, Apr 6, 2020 at 1:14 PM GS [e-mail] wrote:

@shouryasimha https://github.com/shouryasimha also why is it
yolov3.backup? Backup shouldn’t be a file extension and should be a folder.
Like this yolov3/backup

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/pjreddie/darknet/issues/416#issuecomment-609923572,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2FZFJQC55VCLBDR2HKTRLIEXZANCNFSM4EMBKEOQ
.

[USNA2014]

Interesting, didn’t know that. Good to know.

[Javaworld22]

Yolov3.backup is the weight corresponding to the latest weights wrt to the epochs , such as if your model stopped at 7123 iteration, using that yolov3.backup you can resume training from there . It’s helpful if you are running your model on colab or so , in case runtime gets disconnected
…
On Mon, Apr 6, 2020 at 1:14 PM GS @.*> wrote: @shouryasimha https://github.com/shouryasimha also why is it yolov3.backup? Backup shouldn’t be a file extension and should be a folder. Like this yolov3/backup — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#416 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALKV2FZFJQC55VCLBDR2HKTRLIEXZANCNFSM4EMBKEOQ .

Please how did you use the doc2unix to solve your problem.
I did this, but its not working.
%cd /content/darknet
!dos2unix ./data/snowman_train.txt
!dos2unix ./data/snowman_test.txt
!dos2unix ./data/coco.data
!dos2unix ./data/yolo.names
!dos2unix ./cfg/yolo1.cfg
!dos2unix ./backup

[shouryasimha]

Please check the address of the backup file .
If you are working on windows , convert the files using dos2unix .
It will be helpful , if you could provide screenshots

On Fri, Apr 24, 2020 at 7:50 AM Javaworld22 [e-mail]
wrote:

Yolov3.backup is the weight corresponding to the latest weights wrt to the
epochs , such as if your model stopped at 7123 iteration, using that
yolov3.backup you can resume training from there . It’s helpful if you are
running your model on colab or so , in case runtime gets disconnected
… <#m_3307301866333555579_>
On Mon, Apr 6, 2020 at 1:14 PM GS @.*> wrote: @shouryasimha
https://github.com/shouryasimha https://github.com/shouryasimha also
why is it yolov3.backup? Backup shouldn’t be a file extension and should be
a folder. Like this yolov3/backup — You are receiving this because you were
mentioned. Reply to this email directly, view it on GitHub <#416 (comment)
https://github.com/pjreddie/darknet/issues/416#issuecomment-609923572>,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2FZFJQC55VCLBDR2HKTRLIEXZANCNFSM4EMBKEOQ
.

Please how did you use the doc2unix to solve your problem.
I did this, but its not working.
%cd /content/darknet
!dos2unix ./data/snowman_train.txt
!dos2unix ./data/snowman_test.txt
!dos2unix ./data/coco.data
!dos2unix ./data/yolo.names
!dos2unix ./cfg/yolo1.cfg
!dos2unix ./backup

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/pjreddie/darknet/issues/416#issuecomment-618964379,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2F2I2LAH2RRXYQEF433ROF4JBANCNFSM4EMBKEOQ
.

[Javaworld22]

Please check the address of the backup file . If you are working on windows , convert the files using dos2unix . It will be helpful , if you could provide screenshots On Fri, Apr 24, 2020 at 7:50 AM Javaworld22 [e-mail] wrote:
…
Yolov3.backup is the weight corresponding to the latest weights wrt to the epochs , such as if your model stopped at 7123 iteration, using that yolov3.backup you can resume training from there . It’s helpful if you are running your model on colab or so , in case runtime gets disconnected … <#m_3307301866333555579_> On Mon, Apr 6, 2020 at 1:14 PM GS @.*> wrote: @shouryasimha https://github.com/shouryasimha https://github.com/shouryasimha also why is it yolov3.backup? Backup shouldn’t be a file extension and should be a folder. Like this yolov3/backup — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#416 (comment) <#416 (comment)>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALKV2FZFJQC55VCLBDR2HKTRLIEXZANCNFSM4EMBKEOQ . Please how did you use the doc2unix to solve your problem. I did this, but its not working. %cd /content/darknet !dos2unix ./data/snowman_train.txt !dos2unix ./data/snowman_test.txt !dos2unix ./data/coco.data !dos2unix ./data/yolo.names !dos2unix ./cfg/yolo1.cfg !dos2unix ./backup — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#416 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALKV2F2I2LAH2RRXYQEF433ROF4JBANCNFSM4EMBKEOQ .

Hello Sir. I just solved the problem now.
whenever it reaches 100th iteration, it stops. Showing me this.

I solved it by commenting this on my Google colab notebook

"!rm /content/darknet/backup -r"
"!ln -s /content/drive/'My Drive'/YOLOv3_weight/backup /content/darknet"

This does the work of relocating my original folder to "/content/drive/'My Drive'/YOLOv3_weight/backup", hence rendering "/content/darknet/backup" invalid

I'm working on macos

[temps1101]

You should run with sudo
Run the command like this.
./darknet detector train cfg/obj.names cfg/yolo-obj.cfg\

this worked for me(ubuntu16).

[shouryasimha]

First install dos2unix
Please find my github repo , and you can your awnser there .

On Fri, Apr 24, 2020 at 11:37 PM temps1101 [e-mail] wrote:

You should run with sudo
Run the command like this.
./darknet detector train cfg/obj.names cfg/yolo-obj.cfg

this worked for me(ubuntu16).

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/pjreddie/darknet/issues/416#issuecomment-619313529,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2F4UBMHLHWPEXTMCWPTROJLGJANCNFSM4EMBKEOQ
.

[FireHead90544]

I ran into the same problem. go to your /cfg/yolo.data file and make sure you have a line that says backup = backup

(not backup = backup/) the code will append "/theweightsfile.weights" to this directory and you'll have double slashes.

Good luck.

[FireHead90544]

I do not have any such file.

[FireHead90544]

@GLorieul Hi, I'm trying to run my model on google colab and gave by a backup path to a folder that I created in google colab but the model stops abruptly at 100th iteration.

Please help me out to solve this issue as I have been stuck with this issue for a week.

I am running on Colab too and facing some issues.

[FireHead90544]

I ran into the same problem. go to your /cfg/yolo.data file and make sure you have a line that says backup = backup
(not backup = backup/) the code will append "/theweightsfile.weights" to this directory and you'll have double slashes.
Good luck.

i got the same issue and was about to try your advise but i see theres no yolo.data file in cfg folder .

Yes.

[sj-on]

I am training my yolo network on an ubuntu machine, and the training seems to go well for the first 100 steps after which it gets aborted with the error: "Couldn't open file:/backup//yolo-obj_100.weights"

Here is the terminal output for reference:

ubuntu@ip-172-31-40-216:~/darknet$ ./darknet detector train cfg/obj.names cfg/yolo-obj.cfg darknet19_448.conv.23
train: Using default 'data/train.list'
backup: Using default '/backup/'
yolo-obj
layer filters size input output
0 conv 32 3 x 3 / 1 416 x 416 x 3 -> 416 x 416 x 32
1 max 2 x 2 / 2 416 x 416 x 32 -> 208 x 208 x 32
2 conv 64 3 x 3 / 1 208 x 208 x 32 -> 208 x 208 x 64
3 max 2 x 2 / 2 208 x 208 x 64 -> 104 x 104 x 64
4 conv 128 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 128
5 conv 64 1 x 1 / 1 104 x 104 x 128 -> 104 x 104 x 64
6 conv 128 3 x 3 / 1 104 x 104 x 64 -> 104 x 104 x 128
7 max 2 x 2 / 2 104 x 104 x 128 -> 52 x 52 x 128
8 conv 256 3 x 3 / 1 52 x 52 x 128 -> 52 x 52 x 256
9 conv 128 1 x 1 / 1 52 x 52 x 256 -> 52 x 52 x 128
10 conv 256 3 x 3 / 1 52 x 52 x 128 -> 52 x 52 x 256
11 max 2 x 2 / 2 52 x 52 x 256 -> 26 x 26 x 256
12 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
13 conv 256 1 x 1 / 1 26 x 26 x 512 -> 26 x 26 x 256
14 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
15 conv 256 1 x 1 / 1 26 x 26 x 512 -> 26 x 26 x 256
16 conv 512 3 x 3 / 1 26 x 26 x 256 -> 26 x 26 x 512
17 max 2 x 2 / 2 26 x 26 x 512 -> 13 x 13 x 512
18 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
19 conv 512 1 x 1 / 1 13 x 13 x1024 -> 13 x 13 x 512
20 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
21 conv 512 1 x 1 / 1 13 x 13 x1024 -> 13 x 13 x 512
22 conv 1024 3 x 3 / 1 13 x 13 x 512 -> 13 x 13 x1024
23 conv 1024 3 x 3 / 1 13 x 13 x1024 -> 13 x 13 x1024
24 conv 1024 3 x 3 / 1 13 x 13 x1024 -> 13 x 13 x1024
25 route 16
26 reorg / 2 26 x 26 x 512 -> 13 x 13 x2048
27 route 26 24
28 conv 35 3 x 3 / 1 13 x 13 x3072 -> 13 x 13 x 35
29 conv 35 1 x 1 / 1 13 x 13 x 35 -> 13 x 13 x 35
30 detection
Loading weights from darknet19_448.conv.23...
seen 32
Done!
Learning Rate: 0.0001, Momentum: 0.9, Decay: 0.0005
Loaded: 1.204687 seconds
Region Avg IOU: 0.327445, Class: 0.507517, Obj: 0.602026, No Obj: 0.534384, Avg Recall: 0.125000, count: 8
Region Avg IOU: 0.373777, Class: 0.478211, Obj: 0.561162, No Obj: 0.535185, Avg Recall: 0.375000, count: 8
Region Avg IOU: 0.214813, Class: 0.614562, Obj: 0.524953, No Obj: 0.534205, Avg Recall: 0.000000, count: 8
Region Avg IOU: 0.314540, Class: 0.647449, Obj: 0.504414, No Obj: 0.535127, Avg Recall: 0.250000, count: 8
Region Avg IOU: 0.286844, Class: 0.568035, Obj: 0.508885, No Obj: 0.533469, Avg Recall: 0.250000, count: 8
Region Avg IOU: 0.214517, Class: 0.626371, Obj: 0.443656, No Obj: 0.534725, Avg Recall: 0.000000, count: 8
.
.
.
.
Region Avg IOU: 0.660878, Class: 0.873684, Obj: 0.239481, No Obj: 0.060791, Avg Recall: 1.000000, count: 8
Region Avg IOU: 0.702378, Class: 0.806532, Obj: 0.313847, No Obj: 0.060674, Avg Recall: 1.000000, count: 8
100: 0.368687, 0.404861 avg, 0.001000 rate, 7.345735 seconds, 6400 images
Saving weights to /backup//yolo-obj_100.weights
Couldn't open file: /backup//yolo-obj_100.weights

Why is this happening? I have the backup folder created at the same directory level as the darknet exe, so, darknet/backup, and the folder is empty.

Any guidance will be very useful

I had one problem too:

And I do not have an exact solution.

[sj-on]

I am running on Google Colab.

[shouryasimha]

Hello Sjon,
If you could describe your problem more, I could help you. if you are working on colab convert your files using !dos2unix. Please check out my GitHub repo https://github.com/shouryasimha/darknet. Please read the read.me .P.S: I was working windows and google colab.

[sj-on]

Hello Sjon,
If you could describe your problem more, I could help you. if you are working on colab convert your files using !dos2unix. Please check out my GitHub repo https://github.com/shouryasimha/darknet. Please read the read.me .P.S: I was working windows and google colab.

[shouryasimha]

@sj-on did you try converting your .data file to dos2unix? This should be done if you have created the file on windows and running on Google colab as colab is built on Linux.

[sj-on]

@sj-on did you try converting your .data file to dos2unix? This should be done if you have created the file on windows and running on Google colab as colab is built on Linux.

Well, I haven't till yet. But my problem did get solve. I don't know how. But next time, if something like this happens, I will try following your method.

[Gujar-Shubham]

I could not find yolo.data file inside cfg

[shouryasimha]

The .data file is where you mention num_class and path to train and
test.txt

On Thu, Jul 16, 2020 at 4:19 PM Shubham arun gujar [e-mail]
wrote:

I could not find yolo.data file inside cfg

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/pjreddie/darknet/issues/416#issuecomment-659648815,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ALKV2F5NBIASNW2HIGVTUYLR35ODNANCNFSM4EMBKEOQ
.

[kushagra522]

I'm using Google Colab.

Saving weights to /content/drive/MyDrive/yolov3/backup/yolov3-custom_last.weights
Couldn't open file: /content/drive/MyDrive/yolov3/backup/yolov3-custom_last.weights

In my darknet/data/obj.data I have saved;
backup = /content/drive/My Drive/yolov3/backup
Please help.
P.S it is saving to MyDrive whereas I'm saving it to My Drive. Does this space affects?

[sj-on]

I'm using Google Colab.

Saving weights to /content/drive/MyDrive/yolov3/backup/yolov3-custom_last.weights
Couldn't open file: /content/drive/MyDrive/yolov3/backup/yolov3-custom_last.weights

In my darknet/data/obj.data I have saved;
backup = /content/drive/My Drive/yolov3/backup
Please help.
P.S it is saving to MyDrive whereas I'm saving it to My Drive. Does this space affects?

Yes, it does. The space changes the whole path. Check whether you are able to find those .weight files and let me know. It is highly unlikely that you will find those files, but if you do, you might be just lucky.

[shouryasimha]

try to create a symbolic link for your path. You can also google yolo on
google colab there's a youtube video that explains everything from start to
end. This could be really helpful or please check out my repo for the
steps. I hope you figure it out!

[RezaSajadian]

Runnin on PowerPC architecture, I was receiving the message as:
Couldn't find yolov3.backup inside the backup folder!
Just manually created the yolov3.backup, to let the process start, and solved the problem this way, and the same file is being overwritten by the further running of the training as the draft of the weights.