Yolov3: Target classes exceed model classes

[joehoeller]

Created custom files for training:
(4 + 1 + 13) * 3 = 54
13 classes, 54 filters

*.names has 13 names in it
*.cfg was converted properly w 13 for classes, 54 for filters in all 3 yolo blocks

yolov3/utils/utils.py", line 451, in build_targets
assert c.max() <= model.nc, 'Target classes exceed model classes'
AssertionError: Target classes exceed model classes

[glenn-jocher]

@joehoeller hey bud. This means that you've supplied class numbers in your labels that exceed the class count you specified in your *.data and *.cfg files. Classes are zero indexed, so for example if you specify classes=3 in *.data and *.cfg, your labels may only have classes 0, 1 and 2 present.

[joehoeller]

So from what I can tell, my labels say 1,2,3 for classes. I can’t seem to
find any other numbers.

Should I just modify my classes to 3 then?

[glenn-jocher]

@joehoeller your *.cfg and *.data simply need to match up in classes. It's ok to use the default yolov3-spp.cfg for example with 80 classes and supply labels that only go up to 15 classes (the remaining 65 spots on the output vectors will go unused, but it won't really hurt your training).

If you modify your classes to 3, then your labels must either be 0, 1, or 2 (zero indexed).

[FranciscoReveriano]

Were you able to solve this? I am starting to think that you might have reference it wrongly in the .data part. Assuming your .cfg file is properly formatted and being called properly.

[joehoeller]

I updated Dark Chocolate, the COCOJSON -> Darknet converter, so now classes are 0 indexed: https://github.com/joehoeller/Dark-Chocolate, and outputs:
0 0.2125 0.369140625 0.0578125 0.232421875

My *.names file has 13 classes (the numbers are there as example, not actually in the file):

0. person
1. bicycle
2. car
3. motorcycle
4. airplane
5. bus
6. train
7. truck
8. boat
9. traffic light
10. fire hydrant
11. stop sign
12. parking meter

The *.data files has this:

classes=13
train=training_img_paths.txt
valid=training_img_paths.txt
names=data/training.names
backup=backup/
eval=coco

I tried the yolov3-spp.cfg as is and got same error as the copy I modified with all 3 [yolo] layers and the conv above it:
(4 + 1 + 13) * 3 = 54

[convolutional]
size=1
stride=1
pad=1
filters=54
activation=linear

[yolo]
mask = 0,1,2
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=13
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
random=1

The error still persists:

Traceback (most recent call last):
  File "train.py", line 450, in <module>
    train()  # train normally
  File "train.py", line 276, in train
    loss, loss_items = compute_loss(pred, targets, model)
  File /yolov3/utils/utils.py", line 333, in compute_loss
    tcls, tbox, indices, anchor_vec = build_targets(model, targets)
  File "/yolov3/utils/utils.py", line 451, in build_targets
    assert c.max() <= model.nc, 'Target classes exceed model classes'
AssertionError: Target classes exceed model classes

[glenn-jocher]

@joehoeller that all looks right. Maybe I should update the error message to output more detail, hold on. Ok I've updated the assert statement in https://github.com/ultralytics/yolov3/commit/0dd0fa7938f605833c85084f7f184445ca0b6fee

Now it should give you more specific information. Can you git pull and try again?

[joehoeller]

@joehoeller that all looks right. Maybe I should update the error message to output more detail, hold on. Ok I've updated the assert statement in 0dd0fa7

Now it should give you more specific information. Can you git pull and try again?

Ok, nice!!! Now I know what the problem is:

AssertionError: Model accepts 13 classes labeled from 0-12, however you labelled a class 17. See https://github.com/ultralytics/yolov3/wiki/Train-Custom-Data

_Let me go fix and I'll report back ;)_

[joehoeller]

I guess it is training, finally:

How long do you think it'll take for 17 classes on a RTX 2080 Ti, using Yolov3-SPP?

[glenn-jocher]

@joehoeller if your terminal window is too short output from tqdm progress bar will wrap.

The images look like your labels need to offset the box centers by box width / 2 and box height / 2

[joehoeller]

No prob, I am on it -> _"...box width / 2 and box height / 2"_, thnx for the tip!
Lastly, the img size says 416 as it's training, but in my *.cfg my img sizes are:

[net]
# Testing
# batch=1
# subdivisions=1
# Training
batch=16
subdivisions=16

width=640
height=512

channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1

[joehoeller]

I adjusted the math, does this look right to you? Darknet/Yolo is all 0.0 to 1.0 values.
Updated vals:

2 0.6578125 0.685546875 0.20625 0.23828125
2 0.81640625 0.6884765625 0.1375 0.150390625
2 0.5625 0.71484375 0.075 0.0546875
2 0.60078125 0.7138671875 0.04375 0.06640625

Previous vals:

2 0.315625 0.37109375 0.20625 0.23828125
2 0.6328125 0.376953125 0.1375 0.150390625
2 0.125 0.4296875 0.075 0.0546875
2 0.2015625 0.427734375 0.04375 0.06640625

[glenn-jocher]

@joehoeller the image size information in the cfg is not used. All of the files (train.py, detect.py, test.py) use the --img-size argument, which is 416 default, and must be a 32-multiple.

Its hard to tell by eye if the new vals are correct, you just need to look at your test_batch0.jpg etc. to see that they overlay your objects correctly.

[joehoeller]

Got it. thnx!!

On Tue, Dec 3, 2019 at 6:22 PM Glenn Jocher [e-mail]
wrote:

@joehoeller https://github.com/joehoeller the image size information in
the cfg is not used. All of the files (train.py, detect.py, test.py) use
the --img-size argument, which is 416 default, and must be a 32-multiple.

Its hard to tell by eye if the new vals are correct, you just need to look
at your test_batch0.jpg etc. to see that they overlay your objects
correctly.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHCKTYGG6ZPU5OLVBYDQW3Z3BA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEF3I5UA#issuecomment-561417936,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHD6UIMISJ5GKEPQUK3QW3Z3BANCNFSM4JUOPTIQ
.

[glenn-jocher]

@joehoeller yeah that looks a lot better. This happens because COCO box origins are in a corner (top left I think), while darknet format has origins in the center.

Note that the overlays you see are the labels, not the predicted boxes. Now you want to let it train for a few hours or a day or so and check your results.png.

So the best way to do it is to note the epoch your validation losses (on the bottom row) start increasing, and then restart your training again from zero to that --epochs number, to lock in the LR drops that are programmed at 80% and 90% of --epochs.

[joehoeller]

Cool. Thanks!! 👌🏽
Yeah, I calculated it’s another 18h @4min per epoch (277 epoch i think it
was).

On Tue, Dec 3, 2019 at 7:14 PM Glenn Jocher [e-mail]
wrote:

@joehoeller https://github.com/joehoeller yeah that looks a lot better.
This happens because COCO box origins are in a corner (top left I think),
while darknet format has origins in the center.

Note that the overlays you see are the labels, not the predicted boxes.
Now you want to let it train for a few hours or a day or so and check your
results.png.

So the best way to do it is to note the epoch your validation losses (on
the bottom row) start increasing, and then restart your training to that
--epochs number (to lock in the LR drops that are programmed at 80% and
90% of --epochs.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHH6JTB24ODXZ4BIIQLQW376NA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEF3L4QY#issuecomment-561430083,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHFCGC44TF5V5TQZAG3QW376NANCNFSM4JUOPTIQ
.

[joehoeller]

@glenn-jocher 4 Ques:

Q1. Is there a way to stop and then resume training?

Q2. You said, "note the epoch your validation losses (on the bottom row) start increasing, ", but this is all I get back:

Q3. When it is done, how to I access metrics like False Positive, False Negative, mAP etc? (Run the cmd for the utils or...?)

Q4. In the image above, is it really producing a mAP score of 0.841 by epoch 109, or is it overfitting?

[joehoeller]

UPDATE: I fixed the PR and updated the math, the COCO JSON -> Darknet conversion tool (Dark Chocolate) works now: https://github.com/joehoeller/Dark-Chocolate/issues/2

[joehoeller]

@glenn-jocher plz see 4 ques above

[FranciscoReveriano]

UPDATE: I fixed the PR and updated the math, the COCO JSON -> Darknet conversion tool (Dark Chocolate) works now: joehoeller/Dark-Chocolate#2

I been super busy with finals and final projects at Duke. So haven't been able to work on this much. But let me know later if you need a code review on Dark Chocolate.

[joehoeller]

Yes plz. Go ahead and review even you can. Good luck w finals!!

On Wed, Dec 4, 2019 at 10:29 PM Francisco Reveriano <
[email protected]> wrote:

UPDATE: I fixed the PR and updated the math, the COCO JSON -> Darknet
conversion tool (Dark Chocolate) works now: joehoeller/Dark-Chocolate#2
https://github.com/joehoeller/Dark-Chocolate/issues/2

I been super busy with finals and final projects at Duke. So haven't been
able to work on this much. But let me know later if you need a code review
on Dark Chocolate.

—
You are receiving this because you modified the open/close state.

Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHG3TGXFNVSR3Q5JKQ3QXB7R7A5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEF7OEUI#issuecomment-561963601,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHFQYH5LJXWNPRV4XL3QXB7R7ANCNFSM4JUOPTIQ
.

[joehoeller]

@glenn-jocher Please provide feedback - >

Done training and got great results, mAP 0.96, which (I think) is competition grade.
My test results after running python3 detect.py --weights weights/last.pt --data data/custom.data --cfg cfg/yolov3-spp-r.cfg --img-size 640 (notice _-r_ is my version of spp.cfg) produced very accurate detection in image set.

So now I am wondering about the following:

1. Where is the final training loss and a graph showing training and validation loss over the epochs?

2. How can I get qualitative detection output images on validation set showing true positive
   detections, false positives and false negatives (how can i do a validation set)?

3. How do I get a class-wise analysis of precision and recall of detections?

4. What would be the column headers for the results.txt file, I am unsure what numbers represent what(?)
   For example, it just outputs this for last set of values:
   272/272 8.13G 1.25 0.349 0.105 1.7 138 416 0.283 0.961 0.922 0.435 1.28 0.31 0.0511

[FranciscoReveriano]

Yeah. if you can request me to do a code review on github. I will be happy to do that.

[joehoeller]

Yeah. if you can request me to do a code review on github. I will be happy to do that.

I tried, it wont let me. I am not sure why, I'll keep trying.

[glenn-jocher]

@joehoeller @FranciscoReveriano sorry I've been pretty busy lately. To answer your questions:

1. The final results should be in results.txt and results.png.
2. For the class by class results run python3 test.py --cfg ... --weights ... --data ... etc.
3. See 2
4. The column headers are in utils.utils.plot_results()

[joehoeller]

Thanks got it!

Great scores by the way:

• Precision Recall mAP
  F1*
  all 8.86e+03 6.76e+04 0.283 0.961 0.922 0.435
  person 8.86e+03 2.24e+04 0.276 0.945 0.9 0.427
  bicycle 8.86e+03 3.99e+03 0.239 0.971 0.932 0.384
  car 8.86e+03 4.13e+04 0.333 0.966 0.936 0.495

On Fri, Dec 6, 2019 at 2:37 PM Glenn Jocher [e-mail]
wrote:

@joehoeller https://github.com/joehoeller @FranciscoReveriano
https://github.com/FranciscoReveriano sorry I've been pretty busy
lately. To answer your questions:

1. The final results should be in results.txt and results.png.
2. For the class by class results run python3 test.py --cfg ...
   --weights ... --data ... etc.
3. See 2
4. The column headers are in utils.utils.plot_results()

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHHRZQPFMDM4M7RI66DQXKZXXA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGFI5XI#issuecomment-562728669,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHHR6NG7PEPU2YQTDFLQXKZXXANCNFSM4JUOPTIQ
.

[joehoeller]

Is there a way to get a mask r-cnn in Ultralytics? (if so, I'd like to contribute).

[joel5638]

@joehoeller how did you fix the
AssertionError: Model accepts 18 classes labeled from 0-17, however you labelled a class 18" issue

[joehoeller]

@joel5638 - I kept the yolov3-spp.cfg in its default configuration, sans the params at the end and at the beginning of each of the three YOLO layers, which were classes and filters. I used the following calculation to determine the number of filters for each layer, based on number of classes:

(4 + 1 + 17) * 3 = 66
17 classes, 66 filters

Scroll down and see number 5 here:
https://github.com/ultralytics/yolov3/wiki/Train-Custom-Data

[joel5638]

@joehoeller okay i fixed it.. and now shape error. the image size of the .tiff is 640x512. Do i have to change the image size in the cfg accordingly or let it be 412?

[joehoeller]

image size issue, use default or --multi-scale

On Mon, Feb 24, 2020 at 10:41 PM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller okay i fixed it.. and now
shape error. the image size of the .tiff is 640x512. Do i have to change
the image size in the cfg accordingly or let it be 412?
[image: Screenshot from 2020-02-25 10-06-45]
https://user-images.githubusercontent.com/38422983/75215565-2e8f5b00-57b7-11ea-91c5-d8673bd7bf63.png

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHC5WYDYPS6U5675IQDRESOP5A5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM2RA6I#issuecomment-590680185,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHDBJVLBCVZNVY5CVIDRESOP5ANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller okay.. Where do i make the changes to default. So i let the images h and w be 640x512 in the cfg. I updated classes = 18, Filters = 66((4+1+n)3) [n=17 classes]

[joehoeller]

in the cfg file, see the wiki link i sent

On Mon, Feb 24, 2020 at 11:04 PM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller okay.. Where do i make the
changes to default. So i let the images h and w be 640x512 in the cfg. I
updated classes = 18, Filters = 66((4+1+n)3) [n=17 classes]

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHETM7K65PHVRKAMVZDRESRGDA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM2SJWQ#issuecomment-590685402,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHERTOY2NEPAKWGVZ2LRESRGDANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller I have my cfg as

[net]

Testing

batch=1

subdivisions=1

Training

batch=16
subdivisions=16

width=640
height=512

channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1

............
............
[convolutional]
size=1
stride=1
pad=1
filters=66
activation=linear

[yolo]
mask = 0,1,2
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=18
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
random=1

still the same error..

I changed the filters and the classes to [filters=66 , classes=18] in all the layers preceding each of the 3 YOLO layers as the link says. But it doesnt work out

[joehoeller]

Hmm not sure. But maybe try this cfg file:

https://github.com/joehoeller/Object-Detection-on-Thermal-Images/blob/master/cfg/yolov3-spp-r.cfg

On Mon, Feb 24, 2020 at 11:16 PM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller I have my cfg as

[net]
Testing batch=1 subdivisions=1 Training

batch=16
subdivisions=16

width=640 height=512

channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1

still the same error..

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHEGT7REMG7O35MDPBLRESSSFA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM2S7TA#issuecomment-590688204,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHCPU6PX3YO3BHMFOJLRESSSFANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller Ive seen this .cfg but the filters are updated as 69 and not 66. should I go ahead and process this cfg file you've shared?

[joehoeller]

Try it and let me know

On Mon, Feb 24, 2020 at 11:47 PM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller Ive seen this .cfg but the
filters are updated as 69 and not 66. should I go ahead and process this
cfg file you've shared?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHGIU64AGUCCKUJ7IMDRESWHRA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM2UZCA#issuecomment-590695560,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHFB7HAZPCN7QWF73NDRESWHRANCNFSM4JUOPTIQ
.

[joel5638]

okay sure

[joel5638]

@joehoeller Still the same error.

Is it because im training this only on the 8,800 16 bit thermal .tiff images?
Im not using any coco dataset.

I'm only training on 16bit thermal .tiff images

Do I also have to add coco to the training set with the thermal images?

If possible can you send me a path of folders you have placed those coco images to?

In the train folder. I have .tiff images.

[joehoeller]

There’s only 3 blocks you need to edit.

On Tue, Feb 25, 2020 at 1:08 AM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller Still the same error.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHGV7UTXS23CUIVXCLTRES7XNA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM22DCY#issuecomment-590717323,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHBTVB3PZWXQW3RXKG3RES7XNANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller yes ive updated the three blocks of the filters, classes.

Do I have to include coco images in the training set?
Or just the ‘.tiff’ will do?

[joehoeller]

Did you add .tiff to the extensions? Torch doesn’t support .tiff out of the
box.

On Tue, Feb 25, 2020 at 8:10 AM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller yes ive updated the three
blocks of the filters, classes.

Do I have to include coco images in the training set?
Or just the ‘.tiff’ will do?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHC3CKOMJZXKJIERZ53REURHDA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM4DBVQ#issuecomment-590885078,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHE5FGMGRMYQX45ZA6LREURHDANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller ive added .tiff extension in the utils/dataset.py

Im just thinking if i should try training with .jpeg first tomorrow and see if I see the error again.

So one question to you, did u train both the 8bit jpeg and coco data for the Thermal object detection?

[joehoeller]

I believe I had jpg images.

On Tue, Feb 25, 2020 at 9:58 AM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller ive added .tiff extension in
the utils/dataset.py

Im just thinking if i should try training with .jpeg first tomorrow and
see if I see the error again.

So one question to you, did u train both the 8bit jpeg and coco data for
the Thermal object detection?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHF654VRFQ5HABYF5UDREU52TA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM4QOSI#issuecomment-590939977,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHBRSO774GEWCMHB223REU52TANCNFSM4JUOPTIQ
.

[joehoeller]

You just need the right math. Try using stock cfg file and make sure ur
classes & file paths match.

On Tue, Feb 25, 2020 at 10:01 AM joe hoeller [e-mail] wrote:

I believe I had jpg images.

On Tue, Feb 25, 2020 at 9:58 AM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller ive added .tiff extension in
the utils/dataset.py

Im just thinking if i should try training with .jpeg first tomorrow and
see if I see the error again.

So one question to you, did u train both the 8bit jpeg and coco data for
the Thermal object detection?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHF654VRFQ5HABYF5UDREU52TA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM4QOSI#issuecomment-590939977,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHBRSO774GEWCMHB223REU52TANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller okay sure. Will try to fix it tomorrow.

[glenn-jocher]

@joel5638 18 classes means n=18

[joel5638]

@glenn-jocher yeah. I changed the filters and updated n=18.
Which means my filters would be (4+1+18)3 = 69

Ive updated the same in the cfg and tried. Doesnt work.

Some minor change is missing out.

So when you asked me to run ‘python3 train.py —img640.

I got an error stating ‘cannot load ../coco/train2014/...

But im not using any coco dataset. Im only training the model on 8000 thermal images but no RGB.

[glenn-jocher]

@joel5638 the example command I gave you is an example of how to use the --img-size argument. Obviously you apply the argument to your own command.

The image format is irrelevant as long as opencv can open them.

[joehoeller]

@glenn: I thought torch didn’t do .tiff out of the box?

On Tue, Feb 25, 2020 at 1:43 PM Glenn Jocher [e-mail]
wrote:

@joel5638 https://github.com/joel5638 the example command I gave you is
an example of how to use the --img-size argument. Obviously you apply
the argument to your own command.

The image format is irrelevant as long as opencv can open them.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHDU6GLXFKPL6SKRBQLREVYHRA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM5HKPQ#issuecomment-591033662,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHELRMGC7AST42IHUP3REVYHRANCNFSM4JUOPTIQ
.

[glenn-jocher]

@joehoeller cv2 loads images and videos:
https://docs.opencv.org/4.2.0/d4/da8/group__imgcodecs.html#ga288b8b3da0892bd651fce07b3bbd3a56

Currently, the following file formats are supported:

Windows bitmaps - *.bmp, *.dib (always supported)
JPEG files - *.jpeg, *.jpg, *.jpe (see the Note section)
JPEG 2000 files - *.jp2 (see the Note section)
Portable Network Graphics - *.png (see the Note section)
WebP - *.webp (see the Note section)
Portable image format - *.pbm, *.pgm, *.ppm *.pxm, *.pnm (always supported)
PFM files - *.pfm (see the Note section)
Sun rasters - *.sr, *.ras (always supported)
TIFF files - *.tiff, *.tif (see the Note section)
OpenEXR Image files - *.exr (see the Note section)
Radiance HDR - *.hdr, *.pic (always supported)
Raster and Vector geospatial data supported by GDAL (see the Note section)

[joehoeller]

Good to know ;)

On Tue, Feb 25, 2020 at 3:27 PM Glenn Jocher [e-mail]
wrote:

@joehoeller https://github.com/joehoeller cv2 loads images and videos.
From
https://docs.opencv.org/4.2.0/d4/da8/group__imgcodecs.html#ga288b8b3da0892bd651fce07b3bbd3a56

Currently, the following file formats are supported:

Windows bitmaps - *.bmp, *.dib (always supported)
JPEG files - *.jpeg, *.jpg, *.jpe (see the Note section)
JPEG 2000 files - *.jp2 (see the Note section)
Portable Network Graphics - *.png (see the Note section)
WebP - *.webp (see the Note section)
Portable image format - *.pbm, *.pgm, *.ppm *.pxm, *.pnm (always supported)
PFM files - *.pfm (see the Note section)
Sun rasters - *.sr, *.ras (always supported)
TIFF files - *.tiff, *.tif (see the Note section)
OpenEXR Image files - *.exr (see the Note section)
Radiance HDR - *.hdr, *.pic (always supported)
Raster and Vector geospatial data supported by GDAL (see the Note section)

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHHHYSF3DDRF7X6OO3TREWELBA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM5SCTA#issuecomment-591077708,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHDZJ72QJBFGMI7M2UDREWELBANCNFSM4JUOPTIQ
.

[joel5638]

@joehoeller @glenn-jocher Im using @joehoeller 's thermal object detection repository. He mentioned 18 classes and ive added the n=18. so filters would be (18+1+4)x3 = 69 which is constant. but it doesnt seem to work.

Ive also tried 66 as filters. I'm lost and confused what is going wrong :(

[joehoeller]

Post the error from the command line

On Tue, Feb 25, 2020 at 9:48 PM Joel Prabhod [e-mail]
wrote:

@joehoeller https://github.com/joehoeller @glenn-jocher
https://github.com/glenn-jocher Im using @joehoeller
https://github.com/joehoeller 's thermal object detection repository.
He mentioned 18 classes and ive added the n=18. so filters would be
(18+1+4)x3 = 69 which is constant. but it doesnt seem to work.

Ive also tried 66 as filters. I'm lost and confused what is going wrong :(

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHEGLAMEXRCP5FI3VITREXRAFA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM6V5XA#issuecomment-591224540,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHCSLG2JX53YT6MKTQLREXRAFANCNFSM4JUOPTIQ
.

[joel5638]

It is the same error.

RuntimeError: shape '[16, 3, 23, 13, 13]' is invalid for input of size 178464

1. I have resized the .tiff image size to 160x120 and using them.
2. Ive updated the classes as 18 and filters as 69 in yolov3-spp-r.cfg
3. Ive also tried classes as 17 and filters as 66 in yolov3-spp-r.cfg

Nothing seems to work.

[joehoeller]

What does your name and data file look like?

On Tue, Feb 25, 2020 at 10:16 PM Joel Prabhod [e-mail]
wrote:

It is the same error.

RuntimeError: shape '[16, 3, 23, 13, 13]' is invalid for input of size
178464

1. I have resized the .tiff image size to 160x120 and using them.
2. Ive updated the classes as 18 and filters as 69 in yolov3-spp-r.cfg
3. Ive also tried classes as 17 and filters as 66 in yolov3-spp-r.cfg

Nothing seems to work.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ultralytics/yolov3/issues/678?email_source=notifications&email_token=ABHVQHC6C55SKOH6A6HWS73REXULJA5CNFSM4JUOPTI2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEM6XJXQ#issuecomment-591230174,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ABHVQHFMFGEMO7DGT5PQX63REXULJANCNFSM4JUOPTIQ
.

[joel5638]

my custom.data looks like this :
classes=18
train=./data/training_img_paths.txt
valid=./data/training_img_paths.txt
names=data/custom.names
backup=backup/
eval=coco

my custom.names looks like this:

person
bicycle
car
motorcycle
airplane
bus
train
truck
boat
traffic light
fire hydrant
stop sign
parking meter
bench
bird
cat
dog
horse

my training_img_paths.txt looks like this:

./coco/images/train/FLIR_00211.tiff
./coco/images/train/FLIR_06774.tiff
./coco/images/train/FLIR_00075.tiff
./coco/images/train/FLIR_00070.tiff
./coco/images/train/FLIR_03503.tiff
......
......
......

[joel5638]

@joehoeller I just cloned your repository again and just added paths to the FLIR dataset with jpeg thermal images and the cfg you used for training.

I ran this command : python3 train.py --data data/custom.data --cfg cfg/yolov3-spp-r.cfg --weights weights/yolov3-spp.weights

when I Change the filters to 69 and classes to 18 in the cfg file. This is the error

Error : AssertionError: Model accepts 18 classes labeled from 0-17, however you labelled a class 18

But when I Change the filters to 66 and classes to 18 in the cfg file. This is the error.

Error : RuntimeError: shape '[16, 3, 23, 13, 13]' is invalid for input of size 178464

Looks like nothing seems to work out with the cfg file

[joehoeller]

Just for fun, try the stock spp cfg file, & see if it works:
https://github.com/joehoeller/Object-Detection-on-Thermal-Images/blob/master/cfg/yolov3-spp.cfg

[glenn-jocher]

@joehoeller @joel5638 yes you can always use the default yolov3-spp.cfg (with no changes) to train custom datasets with up to 80 classes, like an 18 class dataset. It's not an optimal solution, but it works.

[joel5638]

@joehoeller @glenn-jocher I just tried with stock cfgand i see the same result.

AssertionError: Model accepts 18 classes labeled from 0-17, however you labelled a class 18. See https://github.com/ultralytics/yolov3/wiki/Train-Custom-Data

[glenn-jocher]

@joel5638 if your data is only labeled for classes between 0 and 17 it’s not possible to see that message. I thought you’d already fixed your labels?

[joel5638]

@joehoeller @glenn-jocher Finally the training has started. I deleted all the 8000 images and their labels and only took 10 images with their labels and tried and it worked out. So, I'm assuming that the images in the train folder need their respective labels to train.

[glenn-jocher]

Oh good!

The images used for training should not all need label files. We routinely take custom datasets and add COCO images sans labels for backgrounds. All that’s required is that the background images are listed in the *.txt file along with the rest of the (labeled) images.

[joel5638]

Oh wow. Now I get the picture.
Thank you so much @glenn-jocher and @joehoeller
You’ve been a great and great help in educating me. Love your passion. I’m grateful very much. Thank you both.