simple-faster-rcnn-pytorch

simple-faster-rcnn-pytorch copied to clipboard

I use your repo to train on voc07 with pytorch 1.4.0, but I tried for several times and can only get a map as 63, even cannot reach 65. I...

CloudShen-107

calculating localization loss

2

https://github.com/chenyuntc/simple-faster-rcnn-pytorch/blob/6a538bdd113bab649f388aa48193c1b1ccc48ee6/trainer.py#L257 I find a small issue. why in this line gt_label can be equal to zero? loc_loss /= ((**gt_label >= 0**).sum().float()) # ignore gt_label==-1 for rpn_loss But in Line 254,...

daodaofr

请问，如何获得 每个 类的 AP 呢？

Emma0118

Always crash when training on second epoch

1

16GB RAM RTX2080(6GB) how to solve this? maybe batch train?

PeterBishop0

https://docs-cupy.chainer.org/en/stable/ The page isn’t redirecting properly

QinZhang99

CPU和内存占用

5

你好，我在重新训练时刚开始是正常的，但是为什么随着训练时间的增加，CPU和内存的占用也在增加，我4核CPU，16G的内存大约在一小时左右就会占满，电脑卡死，这是什么原因

xungeer29

ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

4

`Traceback (most recent call last): File "/home/dell/.local/lib/python3.6/site-packages/cupy/__init__.py", line 11, in from cupy import core # NOQA File "/home/dell/.local/lib/python3.6/site-packages/cupy/core/__init__.py", line 1, in from cupy.core import core # NOQA ImportError: libcublas.so.9.0: cannot...

XGuider

/home/user/anaconda3/bin/python /home/user/wangxu_data/code/2-AutoDetect/simple-faster-rcnn-pytorch-master/train.py Traceback (most recent call last): File "/home/user/wangxu_data/code/2-AutoDetect/simple-faster-rcnn-pytorch-master/train.py", line 3, in import cupy as cp File "/home/user/anaconda3/lib/python3.7/site-packages/cupy/__init__.py", line 21, in from cupy import core # NOQA File "/home/user/anaconda3/lib/python3.7/site-packages/cupy/core/__init__.py", line...

bevarb

Hi! @chenyuntc If I set `test_num` in file `config.py` to number smaller than len(test_dataloader),then the code ` if ii == test_num: break` will execute,right? But I Encountered this problem! ...

moFang222

A question about calc_detection_voc_prec_rec function in eval_tools

1

i' m confused about these code calc_detection_voc_prec_rec function following: ``` # VOC evaluation follows integer typed bounding boxes. pred_bbox_l = pred_bbox_l.copy() pred_bbox_l[:, 2:] += 1 gt_bbox_l = gt_bbox_l.copy() gt_bbox_l[:, 2:]...

MHX1203