sfzhang15/ATSS issues and pull requests

#103 - cannot import name '_C' from 'atss_core'

Issue - State: open - Opened by dongyanlin 6 months ago

#102 - 大batch_size和多目标检测的双层for循环导致ATSS算法执行时间长，GPU利用率低，如何提升性能与效率？

Issue - State: open - Opened by chenrui17 about 2 years ago - 3 comments

#101 - C1083: 无法打开包括文件: “cpu/vision.h”: No such file or directory

Issue - State: open - Opened by Deill2 about 2 years ago - 1 comment

#100 - No module named 'atss_core'

Issue - State: open - Opened by xiaopan999 over 2 years ago - 1 comment

#99 - where is the code of lite version for FCOS

Issue - State: open - Opened by Justlovesmile over 2 years ago

#98 - Install Issue, children need your help!

Issue - State: open - Opened by YuYue26 almost 3 years ago

#97 - YOLOv3

Issue - State: open - Opened by Jiangyangya about 3 years ago

#96 - Train on own dataset

Issue - State: open - Opened by sh7jacobi about 3 years ago

#95 - training issues

Issue - State: open - Opened by Lixia1221 about 3 years ago

#94 - The full version of ATSS applied to FCOS

Issue - State: open - Opened by thisisi3 over 3 years ago - 3 comments

#93 - Update loss.py

Pull Request - State: open - Opened by johnson-magic over 3 years ago

#92 - total stride size

Issue - State: open - Opened by SnailForce over 3 years ago

#91 - Can you provide the result of multi-scale training with Resnet50 backbone in 2x or 3x?

Issue - State: closed - Opened by tangjiuqi097 over 3 years ago

#90 - A small bug.

Issue - State: open - Opened by aa1234241 over 3 years ago

#89 - About sigmoid focal loss cpu.

Issue - State: open - Opened by aa1234241 over 3 years ago

#88 - how do I run the setup.py file

Issue - State: open - Opened by xiaosa96 over 3 years ago - 1 comment

#87 - resnet50 is not official version?

Issue - State: closed - Opened by johnson-magic over 3 years ago - 1 comment

#86 - where is the ATSS?

Issue - State: open - Opened by 123wk45678 over 3 years ago - 5 comments

#85 - How can I do batch processing in demo?

Issue - State: open - Opened by HansolEom over 3 years ago

#84 - How to get the model's test-dev result? Is that right if I change or edit some coding lines in test_net.py?

Issue - State: open - Opened by Alan-D-Chen over 3 years ago - 2 comments

#83 - file

Issue - State: open - Opened by Lixia1221 almost 4 years ago - 1 comment

#82 - FileNotFoundError

Issue - State: open - Opened by Lixia1221 almost 4 years ago

#81 - Who has actually trained this model? The effect of my training model is quite different！

Issue - State: closed - Opened by Alan-D-Chen almost 4 years ago - 4 comments

#80 - I need some help,please.

Issue - State: closed - Opened by Alan-D-Chen almost 4 years ago - 1 comment

#79 - _C.ml_nms

Issue - State: closed - Opened by LGang-coder almost 4 years ago

#78 - RuntimeError: cannot perform reduction function min on tensor with no elements because the operation does not have an identity

Issue - State: closed - Opened by Alan-D-Chen almost 4 years ago - 1 comment

#77 - Can atss be used for rotated gt box and anchor？Thanks!

Issue - State: closed - Opened by LoveDLWujing almost 4 years ago - 1 comment

#76 - so slow

Issue - State: closed - Opened by mytk2012 almost 4 years ago - 1 comment

#75 - ATSS+FCOS

Issue - State: closed - Opened by Tangzhaotz almost 4 years ago - 1 comment

#74 - About Centerness test

Issue - State: closed - Opened by jianjiandandande almost 4 years ago - 1 comment

#73 - When I begin to train and get env infomation, I get a bin/python': double free or corruption (!prev) error😭

Issue - State: closed - Opened by yfzhang114 almost 4 years ago - 1 comment

#72 - Running test_ net.py How is the test data loaded?

Issue - State: closed - Opened by Alan-D-Chen almost 4 years ago - 4 comments

#71 - install issue

Issue - State: closed - Opened by Alan-D-Chen about 4 years ago

#70 - How about employing the IoU between prediction box and ground truth

Issue - State: closed - Opened by Xudangliatiger about 4 years ago - 2 comments

#69 - "WorkerId": an object with that name is already defined

Issue - State: closed - Opened by manhongnie about 4 years ago - 2 comments

#68 - Is it possible to use cpu version when inference?

Issue - State: closed - Opened by HansolEom about 4 years ago - 2 comments

#67 - [Question] Is YOLO series applicable?

Issue - State: closed - Opened by CaptainEven about 4 years ago - 3 comments

#66 - 请问数据集怎么配置

Issue - State: closed - Opened by zj717754140 about 4 years ago

#65 - topk设置

Issue - State: closed - Opened by zuop-star about 4 years ago - 1 comment

#64 - 网络模型切换

Issue - State: closed - Opened by SnailForce about 4 years ago - 1 comment

#63 - python setup.py build develop --no-deps,

Issue - State: closed - Opened by Adnan-annan about 4 years ago - 1 comment

#62 - 对于ATSS，是否可以适用于其他keypoint based的网络？

Issue - State: closed - Opened by JoJoliking about 4 years ago - 1 comment

#61 - centerness branch shares convs with cls or reg?

Issue - State: closed - Opened by huanyuxinchen about 4 years ago - 1 comment

#60 - what is your final model used by ATSS in table 8?

Issue - State: closed - Opened by HyogiJeon about 4 years ago - 3 comments

#59 - How to clarify it is a better target assigner rather than a noise added in the label????

Issue - State: closed - Opened by AlexWang1900 about 4 years ago - 3 comments

#58 - How do I learn multiple anchors?

Issue - State: closed - Opened by HansolEom about 4 years ago - 1 comment

#57 - why the result of ATSS_dcnv2_X_101_64x4d_FPN_2x is so bad in my test case?

Issue - State: closed - Opened by ZhengLeon about 4 years ago - 1 comment

#56 - question about table2

Issue - State: closed - Opened by xulingzhi951118 about 4 years ago - 2 comments

#55 - torch.isnan(centerness).any()

Issue - State: closed - Opened by jingziyou about 4 years ago - 1 comment

#54 - where is the code of ATSS specifically?

Issue - State: closed - Opened by liushuai839 about 4 years ago - 1 comment

#53 - Can't I use pytorch 1.5?

Issue - State: closed - Opened by HansolEom about 4 years ago - 7 comments

#52 - About FCOS+ATSS in Table 3

Issue - State: closed - Opened by youngwanLEE about 4 years ago - 2 comments

#51 - about ab and af inconsistency

Issue - State: closed - Opened by zhenghao977 about 4 years ago - 1 comment

#50 - 请问这个只能在linux下运行吗

Issue - State: closed - Opened by lanjunb about 4 years ago - 1 comment

#49 - @thisisi3

Issue - State: closed - Opened by zzw-l about 4 years ago - 1 comment

#48 - Add ATSS to RPN

Issue - State: closed - Opened by thisisi3 about 4 years ago - 2 comments

#47 - error in table6 in ATSS paper

Issue - State: closed - Opened by ritaiask over 4 years ago - 1 comment

#46 - use 'test_net.py' evaluate on 'coco_2017_test_dev'

Issue - State: closed - Opened by SuperTiming over 4 years ago - 2 comments

#45 - ImportError: No module named 'torchvision'

Issue - State: closed - Opened by ddddddhy over 4 years ago - 2 comments

#44 - when will release the code based on detectron2?

Issue - State: open - Opened by TengFeiHan0 over 4 years ago - 4 comments

#43 - inference time

Issue - State: closed - Opened by hzh8311 over 4 years ago - 1 comment

#42 - IOU

Issue - State: closed - Opened by xwshan over 4 years ago - 2 comments

#41 - when use 3 anchors per pixel， ATSS performance become lower。

Issue - State: closed - Opened by lji72 over 4 years ago - 5 comments

#40 - ImportError: cannot import name '_C' from 'atss_core'

Issue - State: closed - Opened by Huanqing1 over 4 years ago - 6 comments

#39 - 与maskrcnn benchmark代码库的兼容性

Issue - State: closed - Opened by NuaaWill over 4 years ago - 1 comment

#38 - Is there an unfair comparsion?

Issue - State: closed - Opened by huizhang0110 over 4 years ago - 3 comments

#37 - What is the difference between 'In GT Box' and 'Spatial constraint'

Issue - State: closed - Opened by leeesangwon over 4 years ago - 4 comments

#36 - Can I use cuda 10 instead of cuda 9?Is there any mistake

Issue - State: closed - Opened by ms-wangjuan over 4 years ago - 2 comments

#35 - How to change the inference threshold?

Issue - State: closed - Opened by feiyuhuahuo over 4 years ago - 1 comment

#34 - class num

Issue - State: closed - Opened by JUNXYU over 4 years ago - 2 comments

#33 - Have you tried ATSS on the original RetinaNet?

Issue - State: closed - Opened by feiyuhuahuo over 4 years ago - 1 comment

#32 - the P and N

Issue - State: closed - Opened by henbucuoshanghai over 4 years ago - 3 comments

#31 - about multi scale training

Issue - State: closed - Opened by lrhjw over 4 years ago - 1 comment

#30 - RetinaNet : Anchor box offsets and GIoU loss

Issue - State: closed - Opened by ajoseph12 over 4 years ago - 2 comments

#29 - hello,can this code run on windows?

Issue - State: closed - Opened by DTW1004 over 4 years ago - 3 comments

#28 - Is ATSS suitable for detecting large aspect ratio objects?

Issue - State: closed - Opened by feiyuhuahuo over 4 years ago - 1 comment

#27 - Will there be a Detectron2 implementation?

Issue - State: closed - Opened by SkeletonOne over 4 years ago - 2 comments

#26 - Centerness

Issue - State: closed - Opened by haibochina over 4 years ago - 2 comments

#25 - I found the atss is base on fpn like model,how can I it use on ssd framework?

Issue - State: closed - Opened by sunzhe09 over 4 years ago - 1 comment

#24 - Appearing "nan" during training

Issue - State: closed - Opened by b03505036 over 4 years ago - 7 comments

#23 - How many epochs did you run?

Issue - State: closed - Opened by HansolEom over 4 years ago - 1 comment

#22 - train with custom datasets

Issue - State: closed - Opened by kaikaizhu over 4 years ago - 2 comments

#21 - where should i use TEST.BBOX_AUG.ENABLED=True to realize multi-scale testing

Issue - State: closed - Opened by daiwc over 4 years ago - 5 comments

#20 - No module named 'atss_core"

Issue - State: closed - Opened by xafarranxera over 4 years ago - 5 comments

#19 - I want use custom dataset

Issue - State: closed - Opened by HansolEom over 4 years ago - 4 comments

#18 - _c error

Issue - State: closed - Opened by HansolEom over 4 years ago - 2 comments

#17 - ImportError: /root/atss/ATSS/atss_core/_C.cpython-37m-x86_64-linux-gnu.so: undefined symbol: __cudaPopCallConfiguration

Issue - State: closed - Opened by Note-Liu over 4 years ago - 2 comments

#16 - tips for finetuning on private dataset based on the pretrained model

Issue - State: closed - Opened by alen-mask over 4 years ago - 5 comments

#15 - UnicodeDecodeError: 'utf-8' codec can't decode byte 0xbd in position 1581: invalid start byte

Issue - State: closed - Opened by pipi2020 almost 5 years ago - 3 comments

#14 - RuntimeError: cuda runtime error (98) : invalid device function at atss_core/csrc/cuda/SigmoidFocalLoss_cuda.cu:139

Issue - State: closed - Opened by hust-kevin almost 5 years ago - 6 comments

#13 - RPN-only models don't have roi_heads

Issue - State: closed - Opened by kaikaizhu almost 5 years ago - 2 comments

#12 - The training process is unstable.

Issue - State: closed - Opened by qianwangn almost 5 years ago - 2 comments

#11 - anchor center is inside gt?

Issue - State: closed - Opened by qianwangn almost 5 years ago - 2 comments

#10 - ImportError: cannot import name '_C'

Issue - State: closed - Opened by KelvinHuang666 almost 5 years ago - 4 comments

#9 - SSD centerness

Issue - State: closed - Opened by yyfyan almost 5 years ago - 4 comments

#8 - Have error when test

Issue - State: closed - Opened by Wangweilai1 almost 5 years ago - 2 comments

#7 - Excuse me, I have some problem when train the model, it seems that the _C.cpython file have some problem, and I have compile the setup.py for several times, it still not working, can you do me a favor? Thank you...

GitHub / sfzhang15/ATSS issues and pull requests