How to Use CSP Version Of Yolov4

[JayantGoel001]

Hii @hhk7734 , I was implementing yolov4-csp but i am facing a issue in Your Latest Version i.e. 3.1.0 . Please Let me know how to overcome this isuue.

KeyError                                  Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/yolov4/common/parser.py in parse_cfg(cfg_path)
    100                 try:
--> 101                     metalayers[layer.name][option] = value
    102                 except KeyError as error:

3 frames
/usr/local/lib/python3.6/dist-packages/yolov4/common/metalayer/yolo_layer.py in __setitem__(self, key, value)
    155             self.__setattr__(f"_{key}", bool(int(value)))
--> 156         elif key in ("mask",):
    157             self.__setattr__(

KeyError: "'new_coords' is not supported"

The above exception was the direct cause of the following exception:

RuntimeError                              Traceback (most recent call last)
<ipython-input-65-18684bd31797> in <module>()
      6 
      7 yolo.config.parse_names("coco.names")
----> 8 yolo.config.parse_cfg("yolov4-csp.cfg")
      9 
     10 yolo.make_model()

/usr/local/lib/python3.6/dist-packages/yolov4/common/config.py in parse_cfg(self, cfg_path)
     64     def parse_cfg(self, cfg_path: str):
     65         (
---> 66             self._metalayers,
     67             self._layer_count,
     68             self._model_name,

/usr/local/lib/python3.6/dist-packages/yolov4/common/parser.py in parse_cfg(cfg_path)
    104                         f"parse_cfg: [{layer.name}] '{option}' is not"
    105                         " supported."
--> 106                     ) from error
    107 
    108     # Build layer

RuntimeError: parse_cfg: [yolo_0] 'new_coords' is not supported.

[hhk7734]

is it yolov4 v3.1.0 ?

> python3 -m pip show yolov4      
Name: yolov4
Version: 3.1.0
Summary: YOLOv4: Optimal Speed and Accuracy of Object Detection
Home-page: https://github.com/hhk7734/tensorflow-yolov4
Author: Hyeonki Hong
Author-email: hhk7734@gmail.com
License: MIT
Location: /home/hhk7734/tensorflow-yolov4/venv/lib/python3.8/site-packages
Requires: numpy
Required-by: 

Example code

from yolov4.tf import YOLOv4

yolo = YOLOv4()
yolo.config.parse_names("test/coco.names")
yolo.config.parse_cfg("config/yolov4-csp.cfg")

yolo.make_model()
yolo.load_weights("yolov4-csp.weights", weights_type="yolo")

yolo.inference("test/test.mp4", is_image=False)

[JayantGoel001]

is it yolov4 v3.1.0 ?

> python3 -m pip show yolov4      
Name: yolov4
Version: 3.1.0
Summary: YOLOv4: Optimal Speed and Accuracy of Object Detection
Home-page: https://github.com/hhk7734/tensorflow-yolov4
Author: Hyeonki Hong
Author-email: hhk7734@gmail.com
License: MIT
Location: /home/hhk7734/tensorflow-yolov4/venv/lib/python3.8/site-packages
Requires: numpy
Required-by: 

Example code

from yolov4.tf import YOLOv4

yolo = YOLOv4()
yolo.config.parse_names("test/coco.names")
yolo.config.parse_cfg("config/yolov4-csp.cfg")

yolo.make_model()
yolo.load_weights("yolov4-csp.weights", weights_type="yolo")

yolo.inference("test/test.mp4", is_image=False)

Yes It is yolov 3.1.0 but it is still new_coords error.

[hhk7734]

can you share your cfg file?

[hhk7734]

where is ther new_coords?

[JayantGoel001]

where is ther new_coords?

Ohh I am Sorry I pasted yolov.cfg file This is the yolov4-csp.cfg file.

[net]
batch=32
width=640
height=640
channels=3

momentum=0.949
learning_rate=0.001
burn_in=1000

max_batches=240000
policy=steps
steps=192000,216000
scales=.1,.1

mosaic=1

#23:104x104 54:52x52 85:26x26 104:13x13 for 416

[convolutional]
batch_normalize=1
filters=32
size=3
stride=1
pad=1
activation=mish

# Downsample

[convolutional]
batch_normalize=1
filters=64
size=3
stride=2
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=32
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

# Downsample

[convolutional]
batch_normalize=1
filters=128
size=3
stride=2
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1,-10

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

# Downsample

[convolutional]
batch_normalize=1
filters=256
size=3
stride=2
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1,-28

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

# Downsample

[convolutional]
batch_normalize=1
filters=512
size=3
stride=2
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1,-28

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

# Downsample

[convolutional]
batch_normalize=1
filters=1024
size=3
stride=2
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish

[shortcut]
from=-3
activation=linear

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1,-16

[convolutional]
batch_normalize=1
filters=1024
size=1
stride=1
pad=1
activation=mish

##########################

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

### SPP ###
[maxpool]
stride=1
size=5

[route]
layers=-2

[maxpool]
stride=1
size=9

[route]
layers=-4

[maxpool]
stride=1
size=13

[route]
layers=-1,-3,-5,-6
### End SPP ###

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=mish

[route]
layers = -1, -13

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[upsample]
stride=2

[route]
layers = 79

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1, -3

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=mish

[route]
layers = -1, -6

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[upsample]
stride=2

[route]
layers = 48

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[route]
layers = -1, -3

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=128
activation=mish

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=128
activation=mish

[route]
layers = -1, -6

[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish

##########################

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=mish

[convolutional]
size=1
stride=1
pad=1
filters=255
activation=logistic

[yolo]
mask = 0,1,2
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
num=9
scale_x_y = 2.0
classes=80
iou_thresh=0.2
iou_loss=ciou
iou_normalizer=0.05
obj_normalizer=4.0
cls_normalizer=0.5
nms_kind=diounms
beta_nms=0.6
new_coords=1

[route]
layers = -4

[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=256
activation=mish

[route]
layers = -1, -20

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=mish

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=mish

[route]
layers = -1,-6

[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=mish

[convolutional]
size=1
stride=1
pad=1
filters=255
activation=logistic

[yolo]
mask = 3,4,5
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
num=9
scale_x_y = 2.0
classes=80
iou_thresh=0.2
iou_loss=ciou
iou_normalizer=0.05
obj_normalizer=1.0
cls_normalizer=0.5
nms_kind=diounms
beta_nms=0.6
new_coords=1

[route]
layers = -4

[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=512
activation=mish

[route]
layers = -1, -49

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[route]
layers = -2

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=mish

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=mish

[route]
layers = -1,-6

[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish

[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=mish

[convolutional]
size=1
stride=1
pad=1
filters=255
activation=logistic

[yolo]
mask = 6,7,8
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
num=9
scale_x_y = 2.0
classes=80
iou_thresh=0.2
iou_loss=ciou
iou_normalizer=0.05
obj_normalizer=0.4
cls_normalizer=0.5
nms_kind=diounms
beta_nms=0.6
new_coords=1

[hhk7734]

hmm..

would you try to reinstall?

python3 -m pip uninstall -y yolov4

python3 -m pip install yolov4 -U --user

[hhk7734]

Do uninstall twice to make sure

[JayantGoel001]

Still Getting The Same Errror

[hhk7734]

It seems to be weired. your script run on /usr/local/lib/python3.6 but yolov4 is installed /root/.local/lib/python3.6 Maybe your ipython is set other python.

[JayantGoel001]

I am running it on Google Colab. Yolov4 model I tried was working Great but why yolov4-csp is giving error?

[hhk7734]

use /usr/bin/python3 instead of python3 and try reinstall

[JayantGoel001]

use /usr/bin/python3 instead of python3 and try reinstall

No Effect Same Error 😥

[hhk7734]

I just tested it in colab, but the same error doesn't come out. Try to factory reset.

[JayantGoel001]

I just tested it in colab, but the same error doesn't come out. Try to factory reset.

Ya It Worked Thanks a lot for helping me out.🙂