WenmuZhou/PAN.pytorch
A unofficial pytorch implementation of PAN(PSENet2): Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network
C++Apache-2.0
Issues
- 0
- 0
关于segmentation_head的输出
#36 opened by xiexinch - 4
重新编译的PSE.so 使用报Segmentation fault
#25 opened by wy51wy - 2
Exception: ZIP entry not valid
#28 opened by mrwu-mac - 1
win 10 下面编译的问题
#39 opened by rolyeMMlin - 19
想问问,作者跑的时候有没有发现边界比较差
#19 opened by GYxiaOH - 0
Finetune checkpoint
#44 opened by adijindal30 - 2
- 4
import error
#31 opened by amandazw - 0
- 3
GPU显存占满,但利用率低
#30 opened by mrwu-mac - 0
Pretrained model
#41 opened by MotiBaadror - 0
post_processing 中的subprocess.call
#40 opened by bowenliu1996 - 1
有没有训练好的模型可以放出来吗
#38 opened by shaohailin - 18
特别密集的文本会有内存溢出的情况
#23 opened by Brave731 - 0
测试精度和召回率都0
#37 opened by zzcqinag - 3
cpp的pse有些问题
#22 opened by xxlxx1 - 3
- 0
Mobilenet model ?
#35 opened by raghavgurbaxani - 0
关于performance中精度和参数的问题
#34 opened by Syencil - 1
predicted similarity vectors.
#32 opened by zishui-wu - 0
关于PAN的损失
#33 opened by ohheysherry66 - 1
similarity vectors是否需要四通道?
#21 opened by xtczq - 6
some problem about pse.cpp
#26 opened by ssxxx1a - 4
请问如何检测curved text呢?
#29 opened by mrwu-mac - 0
KeyError: 'metrics'
#20 opened by muzeshan - 4
不能使用gpu训练
#18 opened by wuxiaolianggit - 6
Weights?
#16 opened by zobeirraisi - 1
请问一下有没有pretrained model可以用来test模型效率呀?
#17 opened by kylechang523 - 2
hello,请问有模型吗?
#15 opened by ssxxx1a - 1
Segmentation fault (core dumped)
#7 opened by cuimiao187561 - 1
- 1
decode_np卡死是怎么回事,具体是pse_cpp函数
#10 opened by hoslo - 1
你好,请问下你pse里的get_points是什么作用
#11 opened by GYxiaOH - 1
导入c++版pse
#13 opened by chen849157649 - 1
from post_processing import decode_np as decode ImportError: cannot import name 'decode_np'
#14 opened by luqiang6q - 7
The loss does not decrease
#3 opened by oysz2016 - 1
- 1
分割网络的输出w,h是原图的1/4,train的label也是原图的1/4吗?那么predict得到bbox的坐标也要*4才是真正的文本区域是吗?
#12 opened by xiaohuihuichao - 0
[疑问]FPEM模块
#8 opened by xiaohuihuichao - 0
fpem 实现方式写错了
#6 opened by xiaomaxiao - 3
Text-line detection
#1 opened - 1
num_samples=0
#4 opened - 1
OCR integration
#2 opened