issues
search
YonghaoHe
/
LFD-A-Light-and-Fast-Detector
LFD is a big update upon LFFD. Generally, LFD is a multi-class object detector characterized by lightweight, low inference latency and superior precision. It is for real-world appilcations.
416
stars
83
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
预训练模型的归一化参数是什么？
#51
jpyjpr
closed
9 months ago
0
Release a CPU version with pytorch or TVM
#50
dexception
opened
2 years ago
1
Output the Quality Focal Loss
#49
ver0z
opened
2 years ago
2
Change from focal loss to quality loss
#48
ver0z
closed
2 years ago
0
Tensorrt error on second and subsequent execute_async calls
#47
kb1ooo
opened
2 years ago
1
Coco: Format error?
#46
RyanCV
opened
2 years ago
1
why trt mode consumed more GPU and MEM space？
#45
chainfitness
opened
2 years ago
2
WIDERFACE_train predict_error
#44
babyblue1334
opened
2 years ago
1
Are negative samples required?
#43
kb1ooo
closed
2 years ago
2
Warning messages that appear when I use the model to make predictions
#42
XiaomuWang
opened
2 years ago
3
当我构建TensorRT时遇到的报错
#41
XiaomuWang
closed
2 years ago
3
请问模型如何导出onnx?
#40
ChungTak
opened
2 years ago
3
errors when run ~/WIDERFACE_train/predict.py
#39
zkxwy1996
opened
2 years ago
1
图片Normalize可省略提升性能
#38
linxiaochunsd
opened
2 years ago
1
Fix README.md
#37
artemiy-shirokov
opened
2 years ago
0
[WARNING: ResNet pretrained weights load] unexpected keys:
#36
XiaomuWang
opened
3 years ago
2
Too many FP results with classification_threshold=0.01
#35
Tsuuuki223
closed
3 years ago
2
How the model predicts in the video？
#34
XiaomuWang
closed
3 years ago
1
Accuracy Discrepancy
#33
prabhd33ps
closed
3 years ago
8
Ubantu , Centos or Windows?
#32
Maping1026
opened
3 years ago
2
min_size_threshold of negative train image
#31
Yaodidida
opened
3 years ago
1
LFD supports variable input resolution training
#30
gouravvemula0
closed
3 years ago
0
OCR with LFD
#29
RamatovInomjon
closed
3 years ago
3
Error: onnx2tensorRT in NVIDIA TX2
#28
HT-Yuan
closed
3 years ago
5
Code runs fine but getting a 'Segmentation fault' when exiting the predict program
#27
Tetsujinfr
opened
3 years ago
2
是否能提供API文档？
#26
ZhengHui-Z
closed
3 years ago
2
Enhance predict.py flexibility and functionalities
#25
Tetsujinfr
closed
3 years ago
1
是否能提供Docker安装镜像。
#24
ZhengHui-Z
closed
3 years ago
0
LFD-resnet
#23
kingloo2014
closed
3 years ago
1
Build/Compile libjpeg-turbo
#22
antoniojmabreu
closed
3 years ago
6
What is the difference between LFFD-Backbone and LFD-Backbone?
#21
LeeKyungwook
closed
3 years ago
1
代码写的相当复杂，可以简化一下
#20
kache1995
closed
3 years ago
2
Can not detect large objects problem
#19
lucasjinreal
opened
3 years ago
16
[BUG] RandomBBoxCropRegionSampler can not handle image size smaller than 512
#18
lucasjinreal
closed
3 years ago
8
How to improve your hard set quality ?
#17
docongminh
closed
3 years ago
6
Custom datset too many positives
#16
lucasjinreal
closed
3 years ago
10
May you release a cpu-based inference version?
#15
on-your-way
closed
3 years ago
7
测试精度问题
#14
Watebear
closed
3 years ago
3
Errors when running prediect.py
#13
on-your-way
closed
3 years ago
3
在3090显卡 CUDA11.0 环境下编译的问题
#12
Watebear
closed
3 years ago
2
[Enhancement] evaluation in every serveral epochs
#11
lucasjinreal
closed
3 years ago
1
Please post some representative detection result images or videos and benchmark results
#10
wwdok
closed
3 years ago
3
DataParallel
#9
wobjtushisui
closed
3 years ago
3
/lfd/lfd/data_pipeline/dataset/utils/turbojpeg.py", line 455, in <module> target_libturbojpeg_path = max(libturtojpeg_path_list) ValueError: max() arg is an empty sequence
#8
lucasjinreal
closed
3 years ago
2
Does gen_neg images important?
#7
lucasjinreal
closed
3 years ago
6
如何训练自己的模型？
#6
toplinuxsir
closed
3 years ago
1
int8 calibrator
#5
dexception
closed
3 years ago
6
MSCOCO dataset is waiting for you.
#4
ashuezy
closed
3 years ago
8
有移动端的性能测试吗？
#3
zhongwenkun886
closed
3 years ago
4
Any model design doc or inference result images?
#2
lucasjinreal
closed
3 years ago
2
Next