Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / msracver/Deformable-ConvNets issues and pull requests

#302 - About the range of offset

Issue - State: open - Opened by Chris-city over 1 year ago - 1 comment

#301 - PyTorch version of deformable convolution does not exist

Issue - State: closed - Opened by ghost over 1 year ago - 1 comment

#299 - Is it useful to use dcn on 1x1conv?

Issue - State: open - Opened by kirkzZ almost 3 years ago - 1 comment

#298 - convert to onnx

Issue - State: open - Opened by ggsddu7 almost 3 years ago

#284 - fix test

Pull Request - State: open - Opened by kinney-sun over 3 years ago

#283 - How to get the sampling points?

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

#282 - How to get the receptive field?

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

#281 - fix error in modulated_deformable_conv function

Pull Request - State: open - Opened by developer0hye over 3 years ago

#280 - Why is the magnitude of mask multiplied by 2?

Issue - State: open - Opened by developer0hye over 3 years ago - 1 comment

#277 - Does anyone sovle the problem of over large offset?

Issue - State: open - Opened by hmjbuaa over 4 years ago - 1 comment

#276 - visualize the deformable convolution of rcnn_dcn

Issue - State: closed - Opened by xuwwwux almost 5 years ago

#275 - DCN for 3D

Issue - State: open - Opened by doramran almost 5 years ago

#274 - The channels' order of offsets layer

Issue - State: open - Opened by whulc about 5 years ago

#272 - compile DCN v2 fatal error: ../linalg.h: No such file or directory

Issue - State: open - Opened by jiashu-zhu about 5 years ago - 4 comments

#271 - Deeplab Training Detail

Issue - State: open - Opened by donhk1011 about 5 years ago

#269 - Can I only use CUDA8.0?

Issue - State: open - Opened by WHY95666 about 5 years ago

#268 - Question about num_deformable_group

Issue - State: closed - Opened by chongruo about 5 years ago - 3 comments

#267 - docs: fix typo

Pull Request - State: open - Opened by Jason-Cooke about 5 years ago

#266 - mxnet.base.MXNetError: Error in operator fpn_p4_sum:

Issue - State: open - Opened by herrickli over 5 years ago - 1 comment

#264 - undefined symbol: _ZN2at7getTypeERKNS_6TensorE

Issue - State: open - Opened by xiaoliqiangxxqzz over 5 years ago - 2 comments

#263 - Update pascal_voc_eval.py

Pull Request - State: open - Opened by light201212 over 5 years ago

#262 - Could you provide results file in json form from DCN v2?

Issue - State: open - Opened by ArchNew over 5 years ago

#261 - ##

Issue - State: closed - Opened by evorta over 5 years ago

#260 - #

Issue - State: closed - Opened by evorta over 5 years ago

#259 - raise MXNetError(py_str(_LIB.MXGetLastError()))

Issue - State: open - Opened by wander1985 over 5 years ago

#258 - keras deformable

Issue - State: open - Opened by ICDI0906 over 5 years ago

#257 - Which version of tensorflow should be used?

Issue - State: closed - Opened by wander1985 over 5 years ago

#256 - What does `[email protected]` mean?

Issue - State: closed - Opened by jiec-msft over 5 years ago - 1 comment

#255 - Is the evaluation correct for pascal segmentation ?

Issue - State: open - Opened by pengwangucla over 5 years ago

#254 - Symbol file to train DCNv2

Issue - State: open - Opened by miraj7789 over 5 years ago

#252 - Confusion about why a single conv layer is able to learn the offsets

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

#249 - Question about dcnv2 ops

Issue - State: open - Opened by henrylee2570 almost 6 years ago

#247 - RCNNLogLoss not decreas

Issue - State: open - Opened by zhengye1995 almost 6 years ago

#246 - Could you explain why mask generated by Sigmoid in mdcn is doubled?

Issue - State: open - Opened by JrPeng almost 6 years ago - 1 comment

#245 - this gemm operation seemes useless

Issue - State: closed - Opened by BIGKnight almost 6 years ago - 1 comment

#243 - loss

Issue - State: closed - Opened by anorthman almost 6 years ago

#242 - In readme, for deeplab, we use 4 GPUs for all experiments

Issue - State: open - Opened by KeyKy almost 6 years ago

#241 - Seems the implementation of offsets is different from the original paper?

Issue - State: open - Opened by SUZhaoyu almost 6 years ago - 1 comment

#240 - why slicing the input data at the batch dimension ?

Issue - State: closed - Opened by Google1234 almost 6 years ago

#239 - No module named skimage.draw

Issue - State: open - Opened by yfnn almost 6 years ago - 1 comment

#238 - RCNNLogLoss not decreasing while traing

Issue - State: open - Opened by yohohohoho almost 6 years ago - 4 comments

#237 - where to embed deform_conv layer?

Issue - State: open - Opened by satchmo233 almost 6 years ago - 2 comments

#236 - A bug？

Issue - State: open - Opened by Google1234 almost 6 years ago - 1 comment

#235 - Question DCNv2 operators

Issue - State: closed - Opened by Google1234 almost 6 years ago - 2 comments

#233 - How to train a classification network using mdcn?

Issue - State: open - Opened by zhanglonghao1992 almost 6 years ago

#232 - the size of feature maps before deform conv

Issue - State: open - Opened by hjy1312 almost 6 years ago - 3 comments

#231 - hi，can you provide the new symbol with Deform v2?

Issue - State: open - Opened by bitwangdan almost 6 years ago

#230 - Question aboubt modulated_deformable_im2col

Issue - State: open - Opened by zhanglonghao1992 almost 6 years ago - 2 comments

#228 - Exception: unknown storage type: -1 when no OHEM

Issue - State: open - Opened by HaydenFaulkner almost 6 years ago - 3 comments

#227 - when i train rfcn_dcn on my own datasets, the detected results is good, but

Issue - State: closed - Opened by bitwangdan almost 6 years ago - 2 comments

#225 - Anyone know how to train a classification network？

Issue - State: open - Opened by zhanglonghao1992 almost 6 years ago

#224 - Compare with multi-grid method

Issue - State: open - Opened by limitime almost 6 years ago

#223 - The deformable convolution layer stride must be set to 1?

Issue - State: open - Opened by Demohai almost 6 years ago

#220 - when i training cicyscades dataset, the error like this

Issue - State: open - Opened by bitwangdan almost 6 years ago - 1 comment

#216 - Something wrong when i change my fpn's backbone to vgg16

Issue - State: open - Opened by Yorionice1 about 6 years ago - 3 comments

#214 - Fpnmodel

Pull Request - State: closed - Opened by imranfateh about 6 years ago - 1 comment

#213 - ImportError: cannot import name bbox_overlaps_cython

Issue - State: open - Opened by lzh19961031 about 6 years ago - 3 comments

#208 - it occurs when train fpn

Issue - State: open - Opened by chanyixialex over 6 years ago - 3 comments

#206 - How to use RoiAlign?

Issue - State: open - Opened by zhoulukuan over 6 years ago - 1 comment

#202 - R-CNN FG Accuracy=0.000000

Issue - State: closed - Opened by SingL3 over 6 years ago - 11 comments

#199 - resnet_v1_101-0000.params

Issue - State: open - Opened by GuoxingYan over 6 years ago - 10 comments

#197 - The code runs, but nothing happens

Issue - State: closed - Opened by Karthik-Suresh93 over 6 years ago

#193 - How to train on custom dataset?

Issue - State: open - Opened by Karthik-Suresh93 over 6 years ago - 3 comments

#192 - MXNet with MKLDNN performs worse on Deformable as compared to without

Issue - State: open - Opened by arunbuduri over 6 years ago - 6 comments

#191 - why there is not vgg16 backbone network?

Issue - State: closed - Opened by zchrissirhcz over 6 years ago

#187 - Error: when training deeplab use CityScapes dataset

Issue - State: open - Opened by Toolazy2cruel over 6 years ago - 5 comments

#182 - support singularity container

Pull Request - State: open - Opened by RudyChin over 6 years ago

#176 - Anyone know if it is good for classification task?

Issue - State: open - Opened by akturtle over 6 years ago - 2 comments

#160 - FPN+DCN on kitti dataset

Issue - State: open - Opened by hedes1992 over 6 years ago - 7 comments

#149 - R-CNN FG Accuracy=0.000000 when training FPN-DCN on coco

Issue - State: closed - Opened by yuzcccc almost 7 years ago - 2 comments

#145 - Merge pull request #1 from msracver/master

Pull Request - State: open - Opened by dreadlord1984 almost 7 years ago

#124 - Can this repo be run on cpu only mxnet?

Issue - State: open - Opened by SunGaofeng almost 7 years ago - 6 comments

#119 - hight-head rcnn on caffe

Pull Request - State: open - Opened by chengshuai almost 7 years ago - 6 comments

#108 - train log seems good, but testing result is very bad

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

#100 - MXNET version unsupprted

Issue - State: closed - Opened by thusiyuan about 7 years ago - 1 comment

#99 - Need res-50 or res-34 version of "rfcn/symbols/resnet_v1_101_rfcn.py"

Issue - State: closed - Opened by arunbuduri about 7 years ago - 15 comments

#97 - ValueError:zero-size array to reduction operation minimum which has no identity

Issue - State: closed - Opened by Mendel1 about 7 years ago - 2 comments

#96 - Python symbol for other ImageNet models

Issue - State: open - Opened by HashSolo about 7 years ago - 1 comment

#94 - how to test a batch in demo

Issue - State: open - Opened by brickjava about 7 years ago