simple-faster-rcnn-pytorch

simple-faster-rcnn-pytorch copied to clipboard

Loss leaps after every epoch

3

YangYangGirl

Hi to everyone, I encounter a strange issue. Due to some reason, I am not able to use visdom on my PC so I was trying to save the images...

ZichengDuan

HI,How to achieve batch_size>1 or use multi_gpu，Can you give me an idea?

4

VikingKang

CUDA_ERROR_INVALID_HANDLE: invalid resource handle occured when using another gpu instead of the default card(0)

1

My server has multi gpus, when I change '.cuda()' to '.cuda(1)', exception occured: cupy.cuda.driver.CUDADriverError: CUDA_ERROR_INVALID_HANDLE: invalid resource handle I also add 'os.environ["CUDA_VISIBLE_DEVICES"] = 1', but it doesn't work. Can anyone...

chrisway613

Can I use another GPU on my computer?

2

I want to use another GPU on my computer, but when I change all .cuda() to .cuda(2) this error occured: `cupy.cuda.driver.CUDADriverError: CUDA_ERROR_INVALID_HANDLE: invalid resource handle` What can I do to...

Flawless1202

Hi, I read about your code in https://github.com/chenyuntc/simple-faster-rcnn-pytorch/blob/master/utils/eval_tool.py#L200 which is as below: selec = np.zeros(gt_bbox_l.shape[0], dtype=bool) for gt_idx in gt_index: if gt_idx >= 0: if gt_difficult_l[gt_idx]: match[l].append(-1) else: if not...

300LiterPropofol

The problem in train on my own voc data

7

Hi, I use my data like voc dataset in this project.and I have met some problem. First, I replace the offical voc data(20 classes) by my own data(5 classes). and...

LDoubleZhi

请教下rool pooling过程

1

 上图是我保存rool pooling池化过程中第0个通道的特征图，所有的rois坐标和第0个rois池化后的结果    shang上面三个图分别是特征图，rois和池化后保存的结果。 第一个问题：我从第0个rois得到缩小16倍的位置在7,8,9,10列。请问这个位置应该是四舍五入还是取整？ 第二个问题：假设在四舍五入的情况下，我得到xy的起终点坐标为[21, 20]->[27, 31]（注意excel以1开始）.得到对下面部分的最大池化  得到12*7的大小，我们要池化为7*7。请问您是怎么对着12划分的呢？比如池化后的第一行，对应特征图的那几行呢？由于rol-pooling是cupy写的，有点看不懂。请问有木有一个公式？

Yuesheng321

Hello there. What is the numpy version of the study? My computer always prompts that numpy（ numpy-1.18.1） is missing various libraries. such as: `Traceback (most recent call last): File "D:/a_zhy/object_detection/simple-faster-rcnn-pytorch-master/train.py",...

COST-97

so this function is to move a given bbox (src_bbox) to offset (loc) which is given as (dy,dx,dh,dw) src_height = src_bbox[:, 2] - src_bbox[:,0] src_width = src_bbox[:, 3] - src_bbox[:,...

ooodragon94