Comments (11)
zjZSTU
commented on June 30, 2024
1
Hi, when it prints:
Not supported field ...
It means that the field in the cfg is not taken into account changing the model of the network and therefore the dimension of the weights to load.
Not all cfg are supported, see DarknetParser.cpp
hi @ceccocats , i use yolov4 pretrained weights in YOLOv4 model zoo, there also will appear Not supported info, but following work is fine
$ ./test_yolo4
Not supported field: batch=1
Not supported field: subdivisions=1
Not supported field: momentum=0.949
Not supported field: decay=0.0005
Not supported field: angle=0
Not supported field: saturation = 1.5
Not supported field: exposure = 1.5
Not supported field: hue=.1
Not supported field: learning_rate=0.00261
Not supported field: burn_in=1000
Not supported field: max_batches = 500500
Not supported field: policy=steps
Not supported field: steps=400000,450000
Not supported field: scales=.1,.1
Not supported field: mosaic=1
New NETWORK (tkDNN v0.5, CUDNN v8)
but when i use my own trained weights, it doesn't work
$ ./test_yolo4
Not supported field: batch=1
Not supported field: subdivisions=1
Not supported field: momentum=0.949
Not supported field: decay=0.0005
Not supported field: angle=0
Not supported field: saturation = 1.5
Not supported field: exposure = 1.5
Not supported field: hue=.1
Not supported field: learning_rate=0.00261
Not supported field: burn_in=1000
Not supported field: max_batches = 500500
Not supported field: policy=steps
Not supported field: steps=400000,450000
Not supported field: scales=.1,.1
Not supported field: mosaic=1
New NETWORK (tkDNN v0.5, CUDNN v8)
Reading weights: I=3 O=32 KERNEL=3x3x1
Reading weights: I=32 O=64 KERNEL=3x3x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=64 O=32 KERNEL=1x1x1
Reading weights: I=32 O=64 KERNEL=3x3x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=128 O=64 KERNEL=1x1x1
Reading weights: I=64 O=128 KERNEL=3x3x1
Reading weights: I=128 O=64 KERNEL=1x1x1
Reading weights: I=128 O=64 KERNEL=1x1x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=64 O=64 KERNEL=3x3x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=64 O=64 KERNEL=3x3x1
Reading weights: I=64 O=64 KERNEL=1x1x1
Reading weights: I=128 O=128 KERNEL=1x1x1
Reading weights: I=128 O=256 KERNEL=3x3x1
Reading weights: I=256 O=128 KERNEL=1x1x1
Reading weights: I=256 O=128 KERNEL=1x1x1
...
...
Reading weights: I=128 O=256 KERNEL=3x3x1
Reading weights: I=256 O=255 KERNEL=1x1x1
Error reading file yolo4/layers/c138.bin with n of float: 65280 seek: 0 size: 261120
/home/user/software/tkDNN/src/utils.cpp:58
Aborting...
thank you for you help
from tkdnn.
ceccocats
commented on June 30, 2024
Hi, when it prints:
Not supported field ...
It means that the field in the cfg is not taken into account changing the model of the network and therefore the dimension of the weights to load.
Not all cfg are supported, see DarknetParser.cpp
from tkdnn.
ASONG0506
commented on June 30, 2024
Hi, I met the similar problem.
I trained my yolov4 model using my own dataset with 10 class, the I converted it using the this repo and no error occured with these infomation:
......
n: 136, type 0
Convolutional
weights: 32768, biases: 128, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c136.bin
n: 137, type 0
Convolutional
weights: 294912, biases: 256, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c137.bin
n: 138, type 0
Convolutional
weights: 11520, biases: 45, batch_normalize: 0, groups: 1
write binary ../output_bdd100k/c138.bin
n: 139, type 27
export YOLO
mask: 3
biases: 18
mask 0.000000
mask 1.000000
mask 2.000000
anchor 12.000000
anchor 16.000000
anchor 19.000000
anchor 36.000000
anchor 40.000000
anchor 28.000000
anchor 36.000000
anchor 75.000000
anchor 76.000000
anchor 55.000000
anchor 72.000000
anchor 146.000000
anchor 142.000000
anchor 110.000000
anchor 192.000000
anchor 243.000000
anchor 459.000000
anchor 401.000000
write binary ../output_bdd100k/g139.bin
n: 140, type 9
export ROUTE
n: 141, type 0
Convolutional
weights: 294912, biases: 256, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c141.bin
n: 142, type 9
export ROUTE
n: 143, type 0
Convolutional
weights: 131072, biases: 256, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c143.bin
n: 144, type 0
Convolutional
weights: 1179648, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c144.bin
n: 145, type 0
Convolutional
weights: 131072, biases: 256, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c145.bin
n: 146, type 0
Convolutional
weights: 1179648, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c146.bin
n: 147, type 0
Convolutional
weights: 131072, biases: 256, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c147.bin
n: 148, type 0
Convolutional
weights: 1179648, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c148.bin
n: 149, type 0
Convolutional
weights: 23040, biases: 45, batch_normalize: 0, groups: 1
write binary ../output_bdd100k/c149.bin
n: 150, type 27
export YOLO
mask: 3
biases: 18
mask 3.000000
mask 4.000000
mask 5.000000
anchor 12.000000
anchor 16.000000
anchor 19.000000
anchor 36.000000
anchor 40.000000
anchor 28.000000
anchor 36.000000
anchor 75.000000
anchor 76.000000
anchor 55.000000
anchor 72.000000
anchor 146.000000
anchor 142.000000
anchor 110.000000
anchor 192.000000
anchor 243.000000
anchor 459.000000
anchor 401.000000
write binary ../output_bdd100k/g150.bin
n: 151, type 9
export ROUTE
n: 152, type 0
Convolutional
weights: 1179648, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c152.bin
n: 153, type 9
export ROUTE
n: 154, type 0
Convolutional
weights: 524288, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c154.bin
n: 155, type 0
Convolutional
weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c155.bin
n: 156, type 0
Convolutional
weights: 524288, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c156.bin
n: 157, type 0
Convolutional
weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c157.bin
n: 158, type 0
Convolutional
weights: 524288, biases: 512, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c158.bin
n: 159, type 0
Convolutional
weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1
write binary ../output_bdd100k/c159.bin
n: 160, type 0
Convolutional
weights: 46080, biases: 45, batch_normalize: 0, groups: 1
write binary ../output_bdd100k/c160.bin
n: 161, type 27
export YOLO
mask: 3
biases: 18
mask 6.000000
mask 7.000000
mask 8.000000
anchor 12.000000
anchor 16.000000
anchor 19.000000
anchor 36.000000
anchor 40.000000
anchor 28.000000
anchor 36.000000
anchor 75.000000
anchor 76.000000
anchor 55.000000
anchor 72.000000
anchor 146.000000
anchor 142.000000
anchor 110.000000
anchor 192.000000
anchor 243.000000
anchor 459.000000
anchor 401.000000
write binary ../output_bdd100k/g161.bin
this is my yolov4.cfg file:
##########################
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky
[convolutional]
size=1
stride=1
pad=1
filters=45
activation=linear
[yolo]
mask = 0,1,2
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
classes=10
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
scale_x_y = 1.2
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5
[route]
layers = -4
[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=256
activation=leaky
[route]
layers = -1, -16
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
[convolutional]
size=1
stride=1
pad=1
filters=45
activation=linear
[yolo]
mask = 3,4,5
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
classes=10
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
scale_x_y = 1.1
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5
[route]
layers = -4
[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=512
activation=leaky
[route]
layers = -1, -37
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky
[convolutional]
size=1
stride=1
pad=1
filters=45
activation=linear
[yolo]
mask = 6,7,8
anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
classes=10
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
random=1
scale_x_y = 1.05
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5
Then I use this repo to generate the .rt file and do inference, when run this commond ./test_yolo4 and error occured as follow:
.......
Reading weights: I=128 O=256 KERNEL=3x3x1
Reading weights: I=256 O=255 KERNEL=1x1x1
Error reading file yolo4/layers/c138.bin with n of float: 65280 seek: 0 size: 261120
/home/xavier/test_ws/tkDNN-master/src/utils.cpp:58
Aborting...
Do you know what's whe wrong with it, thanks!
from tkdnn.
ASONG0506
commented on June 30, 2024
Hi, I met the similar problem.
I trained my yolov4 model using my own dataset with 10 class, the I converted it using the this repo and no error occured with these infomation:...... n: 136, type 0 Convolutional weights: 32768, biases: 128, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c136.bin n: 137, type 0 Convolutional weights: 294912, biases: 256, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c137.bin n: 138, type 0 Convolutional weights: 11520, biases: 45, batch_normalize: 0, groups: 1 write binary ../output_bdd100k/c138.bin n: 139, type 27 export YOLO mask: 3 biases: 18 mask 0.000000 mask 1.000000 mask 2.000000 anchor 12.000000 anchor 16.000000 anchor 19.000000 anchor 36.000000 anchor 40.000000 anchor 28.000000 anchor 36.000000 anchor 75.000000 anchor 76.000000 anchor 55.000000 anchor 72.000000 anchor 146.000000 anchor 142.000000 anchor 110.000000 anchor 192.000000 anchor 243.000000 anchor 459.000000 anchor 401.000000 write binary ../output_bdd100k/g139.bin n: 140, type 9 export ROUTE n: 141, type 0 Convolutional weights: 294912, biases: 256, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c141.bin n: 142, type 9 export ROUTE n: 143, type 0 Convolutional weights: 131072, biases: 256, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c143.bin n: 144, type 0 Convolutional weights: 1179648, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c144.bin n: 145, type 0 Convolutional weights: 131072, biases: 256, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c145.bin n: 146, type 0 Convolutional weights: 1179648, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c146.bin n: 147, type 0 Convolutional weights: 131072, biases: 256, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c147.bin n: 148, type 0 Convolutional weights: 1179648, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c148.bin n: 149, type 0 Convolutional weights: 23040, biases: 45, batch_normalize: 0, groups: 1 write binary ../output_bdd100k/c149.bin n: 150, type 27 export YOLO mask: 3 biases: 18 mask 3.000000 mask 4.000000 mask 5.000000 anchor 12.000000 anchor 16.000000 anchor 19.000000 anchor 36.000000 anchor 40.000000 anchor 28.000000 anchor 36.000000 anchor 75.000000 anchor 76.000000 anchor 55.000000 anchor 72.000000 anchor 146.000000 anchor 142.000000 anchor 110.000000 anchor 192.000000 anchor 243.000000 anchor 459.000000 anchor 401.000000 write binary ../output_bdd100k/g150.bin n: 151, type 9 export ROUTE n: 152, type 0 Convolutional weights: 1179648, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c152.bin n: 153, type 9 export ROUTE n: 154, type 0 Convolutional weights: 524288, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c154.bin n: 155, type 0 Convolutional weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c155.bin n: 156, type 0 Convolutional weights: 524288, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c156.bin n: 157, type 0 Convolutional weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c157.bin n: 158, type 0 Convolutional weights: 524288, biases: 512, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c158.bin n: 159, type 0 Convolutional weights: 4718592, biases: 1024, batch_normalize: 1, groups: 1 write binary ../output_bdd100k/c159.bin n: 160, type 0 Convolutional weights: 46080, biases: 45, batch_normalize: 0, groups: 1 write binary ../output_bdd100k/c160.bin n: 161, type 27 export YOLO mask: 3 biases: 18 mask 6.000000 mask 7.000000 mask 8.000000 anchor 12.000000 anchor 16.000000 anchor 19.000000 anchor 36.000000 anchor 40.000000 anchor 28.000000 anchor 36.000000 anchor 75.000000 anchor 76.000000 anchor 55.000000 anchor 72.000000 anchor 146.000000 anchor 142.000000 anchor 110.000000 anchor 192.000000 anchor 243.000000 anchor 459.000000 anchor 401.000000 write binary ../output_bdd100k/g161.binthis is my yolov4.cfg file:
########################## [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=256 activation=leaky [convolutional] size=1 stride=1 pad=1 filters=45 activation=linear [yolo] mask = 0,1,2 anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401 classes=10 num=9 jitter=.3 ignore_thresh = .7 truth_thresh = 1 scale_x_y = 1.2 iou_thresh=0.213 cls_normalizer=1.0 iou_normalizer=0.07 iou_loss=ciou nms_kind=greedynms beta_nms=0.6 max_delta=5 [route] layers = -4 [convolutional] batch_normalize=1 size=3 stride=2 pad=1 filters=256 activation=leaky [route] layers = -1, -16 [convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky [convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky [convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky [convolutional] size=1 stride=1 pad=1 filters=45 activation=linear [yolo] mask = 3,4,5 anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401 classes=10 num=9 jitter=.3 ignore_thresh = .7 truth_thresh = 1 scale_x_y = 1.1 iou_thresh=0.213 cls_normalizer=1.0 iou_normalizer=0.07 iou_loss=ciou nms_kind=greedynms beta_nms=0.6 max_delta=5 [route] layers = -4 [convolutional] batch_normalize=1 size=3 stride=2 pad=1 filters=512 activation=leaky [route] layers = -1, -37 [convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky [convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky [convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky [convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky [convolutional] size=1 stride=1 pad=1 filters=45 activation=linear [yolo] mask = 6,7,8 anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401 classes=10 num=9 jitter=.3 ignore_thresh = .7 truth_thresh = 1 random=1 scale_x_y = 1.05 iou_thresh=0.213 cls_normalizer=1.0 iou_normalizer=0.07 iou_loss=ciou nms_kind=greedynms beta_nms=0.6 max_delta=5Then I use this repo to generate the .rt file and do inference, when run this commond
./test_yolo4and error occured as follow:....... Reading weights: I=128 O=256 KERNEL=3x3x1 Reading weights: I=256 O=255 KERNEL=1x1x1 Error reading file yolo4/layers/c138.bin with n of float: 65280 seek: 0 size: 261120 /home/xavier/test_ws/tkDNN-master/src/utils.cpp:58 Aborting...Do you know what's whe wrong with it, thanks!
I found the solution, thanks
from tkdnn.
ceccocats
commented on June 30, 2024
Hi, can you share your solution? It can be useful.
It was your mistake? or an error in the code?
from tkdnn.
Kr4is
commented on June 30, 2024
Hi!, my error was that you said.
I had a maxpool layer with maxpool_depth=1 that is not already supported.
I change this layer by:
[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=leaky
As @AlexeyAB says in this issue.
I regenerate the weights, and it works!
Then i think in the previous layer can be an unsupported parameter?
from tkdnn.
ceccocats
commented on June 30, 2024
Yes, darknet Is big and the time is limited.
We decided to implement only the official YOLOs and the ones that needs only minor changes.
from tkdnn.
ASONG0506
commented on June 30, 2024
There is no error in your code, actually I just changed my own cfg file and the default cfg file for parsing the darknet network is : root/tests/darknet/cfg/yolov4.cfg and the corresponding names file is : root/tests/darknet/names/coco.names. when I modified these two files, it worked!
from tkdnn.
mive93
commented on June 30, 2024
Hi @zjZSTU
which cfg are you using in the test?
from tkdnn.
zjZSTU
commented on June 30, 2024
Hi @zjZSTU
which cfg are you using in the test?
hi @mive93 , i have solved my problem. the key is to choose my own .cfg and .name file for custom dataset. refer to #99
from tkdnn.
mive93
commented on June 30, 2024
Yes, exactly 👍
from tkdnn.
Related Issues (20)
- Any updates if it works with new versions of CUDNN? HOT 6
- terminate called after throwing an instance of 'YAML::ParserException' HOT 3
- Build error window HOT 4
- cloud not build cuda engine HOT 2
- Google colab cmake error HOT 1
- not matching Darknet HOT 1
- YOLO v7 Support HOT 7
- How to interface with python Torchrt models? HOT 1
- centernet3d and centertrack3d, too many inferences results showing! HOT 5
- TkDnn vs OpenCv Dnn
- monodepth2 download HOT 3
- Yolo3Detection segfault on cv::resize - Jetpack 5.0 HOT 3
- Problems with custom dataset and tensorrt8
- Can not find a yaml-cpp when build in windows
- Cuda failure: out of memory HOT 2
- Any updates if it works with new versions of cuda 11.8? HOT 3
- I'm missing libdarknetRT.so file after make CMakeLists.txt file
- Significant errors in confidence scores after TensorRT conversion
- Jetson Orin?
- how to run the inference using RT file
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tkdnn.