zhaoweicai / detectron-cascade-rcnn Goto Github PK
View Code? Open in Web Editor NEWCascade R-CNN in Detectron
License: Apache License 2.0
Cascade R-CNN in Detectron
License: Apache License 2.0
大概多久能上传呢?十一之后能上传吗?
Ran 2 tests in 0.726s
OK
I have a custom dataset, Detectron works well, but I switch to your Cascade-RCNN, the loss doesn't decease, and performance is bad, Could you help figure it out? by the way, the gt numbers of my custom dataset is large
The link about the pretrained model seems Invalid
Can you give me some suggestion to solve it?
INFO io.py: 67: Downloading remote file https://s3-us-west-2.amazonaws.com/detectron/ImageNetPretrained/25093814/X-152-32x8d-IN5k.pkl to /tmp/detectron-download-cache/ImageNetPretrained/25093814/X-152-32x8d-IN5k.pkl
Traceback (most recent call last):
File "tools/train_net.py", line 132, in
main()
File "tools/train_net.py", line 101, in main
assert_and_infer_cfg()
File "/home/wrc/CBNet/detectron/core/config.py", line 1127, in assert_and_infer_cfg
cache_cfg_urls()
File "/home/wrc/CBNet/detectron/core/config.py", line 1136, in cache_cfg_urls
__C.TRAIN.WEIGHTS = cache_url(__C.TRAIN.WEIGHTS, __C.DOWNLOAD_CACHE)
File "/home/wrc/CBNet/detectron/utils/io.py", line 68, in cache_url
download_url(url, cache_file_path)
File "/home/wrc/CBNet/detectron/utils/io.py", line 114, in download_url
response = urllib2.urlopen(url)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 154, in urlopen
return opener.open(url, data, timeout)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 435, in open
response = meth(req, response)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 548, in http_response
'http', request, response, code, msg, hdrs)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 473, in error
return self._call_chain(*args)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 407, in _call_chain
result = func(*args)
File "/home/wrc/anaconda3/envs/py27/lib/python2.7/urllib2.py", line 556, in http_error_default
raise HTTPError(req.get_full_url(), code, msg, hdrs, fp)
urllib2.HTTPError: HTTP Error 301: Moved Permanently
train normally
urllib2.HTTPError: HTTP Error 301: Moved Permanently
E.g.:
convert_pkl_to_pb.py does not work for the Detectron-Cascade-RCNN
thx
GPU Types is GTX1080Ti
When performing target detection, the number of different types differs greatly. How can I solve this problem?
INFO infer_simple.py: 133: Processing demo/34501842524_3c858b3080_k.jpg -> demo/output/34501842524_3c858b3080_k.jpg.pdf
INFO infer_simple.py: 141: Inference time: 0.164s
INFO infer_simple.py: 143: | im_detect_bbox: 0.141s
INFO infer_simple.py: 143: | misc_mask: 0.013s
INFO infer_simple.py: 143: | im_detect_mask: 0.006s
INFO infer_simple.py: 143: | misc_bbox: 0.002s
INFO infer_simple.py: 133: Processing demo/19064748793_bb942deea1_k.jpg -> demo/output/19064748793_bb942deea1_k.jpg.pdf
INFO infer_simple.py: 141: Inference time: 0.235s
INFO infer_simple.py: 143: | im_detect_bbox: 0.137s
INFO infer_simple.py: 143: | misc_mask: 0.061s
INFO infer_simple.py: 143: | im_detect_mask: 0.034s
INFO infer_simple.py: 143: | misc_bbox: 0.003s
INFO infer_simple.py: 133: Processing demo/17790319373_bd19b24cfc_k.jpg -> demo/output/17790319373_bd19b24cfc_k.jpg.pdf
INFO infer_simple.py: 141: Inference time: 0.205s
INFO infer_simple.py: 143: | im_detect_bbox: 0.142s
INFO infer_simple.py: 143: | misc_mask: 0.043s
INFO infer_simple.py: 143: | im_detect_mask: 0.017s
INFO infer_simple.py: 143: | misc_bbox: 0.003s
INFO infer_simple.py: 133: Processing demo/16004479832_a748d55f21_k.jpg -> demo/output/16004479832_a748d55f21_k.jpg.pdf
INFO infer_simple.py: 141: Inference time: 0.163s
INFO infer_simple.py: 143: | im_detect_bbox: 0.139s
INFO infer_simple.py: 143: | misc_mask: 0.015s
INFO infer_simple.py: 143: | im_detect_mask: 0.006s
INFO infer_simple.py: 143: | misc_bbox: 0.003s
INFO infer_simple.py: 133: Processing demo/33887522274_eebd074106_k.jpg -> demo/output/33887522274_eebd074106_k.jpg.pdf
INFO infer_simple.py: 141: Inference time: 0.150s
INFO infer_simple.py: 143: | im_detect_bbox: 0.130s
INFO infer_simple.py: 143: | misc_mask: 0.011s
INFO infer_simple.py: 143: | im_detect_mask: 0.006s
INFO infer_simple.py: 143: | misc_bbox: 0.003s
I can't see the files in the output folder under the local demo (I deleted all the files from the original output)
CRITICAL train.py: 98: roi_data_loader failed
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
INFO loader.py: 126: Stopping enqueue thread
When I use my own data to train the model I found the error happened ? I have already modified the lr to a small value but after few hours training ,the error still happened
I use the detectron model and run the e2e_cascade_rcnn_R-50-fpn_1x.yaml
but it only show the MEAN AP
how to show the AP ?
The caffe2 installed can run the official Detectron properly.But when I try to test your code ,this error happend. Can you give me some suggestions?
File "/home/wrc/Detectron-Cascade-RCNN/detectron/utils/env.py", line 71, in get_detectron_ops_lib
('Detectron ops lib not found; make sure that your Caffe2 '
AssertionError: Detectron ops lib not found; make sure that your Caffe2 version includes Detectron module
In your open source code, I didn't find ResNeXt-152. Is it open source too?
I still have a question. Does ‘Cascade R-CNN∗⋆’ in your paper mean ‘Cascade Mask R-CNN’? Are the two the same?
When I train the model with "configs/e2e_cascade_rcnn_R-101-FPN_1x_gn.yaml", it shows such errors:
INFO io.py: 67: Downloading remote file https://s3-us-west-2.amazonaws.com/detectron/ImageNetPretrained/47592356/R-101-GN.pkl to /tmp/detectron-download-cache/ImageNetPretrained/47592356/R-101-GN.pkl
Traceback (most recent call last):
File "tools/train_net.py", line 132, in
main()
File "tools/train_net.py", line 101, in main
assert_and_infer_cfg()
File "/home/test/Detectron-Cascade-RCNN/detectron/core/config.py", line 1127, in assert_and_infer_cfg
cache_cfg_urls()
File "/home/test/Detectron-Cascade-RCNN/detectron/core/config.py", line 1136, in cache_cfg_urls
__C.TRAIN.WEIGHTS = cache_url(__C.TRAIN.WEIGHTS, __C.DOWNLOAD_CACHE)
File "/home/test/Detectron-Cascade-RCNN/detectron/utils/io.py", line 68, in cache_url
download_url(url, cache_file_path)
File "/home/test/Detectron-Cascade-RCNN/detectron/utils/io.py", line 114, in download_url
response = urllib2.urlopen(url)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 154, in urlopen
return opener.open(url, data, timeout)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 435, in open
response = meth(req, response)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 548, in http_response
'http', request, response, code, msg, hdrs)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 473, in error
return self._call_chain(*args)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 407, in _call_chain
result = func(*args)
File "/home/dl/anaconda3/envs/caffe2_py2/lib/python2.7/urllib2.py", line 556, in http_error_default
raise HTTPError(req.get_full_url(), code, msg, hdrs, fp)
urllib2.HTTPError: HTTP Error 404: Not Found
And I try to “wget https://s3-us-west-2.amazonaws.com/detectron/ImageNetPretrained/47592356/R-101-GN.pkl”, it also failured.
Can anyone show me where is the model "https://s3-us-west-2.amazonaws.com/detectron/ImageNetPretrained/47592356/R-101-GN.pkl"?
Hi,
Can you share baseline results for instance segmentation on Cityscapes dataset?
可否更新一下cascade rcnn里面的配置文件?
hi @zhaoweicai
do u do cascaded-heads for mask branch?
hi @zhaoweicai
the perforamce u list on the README
is tested on valset or testdev?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.