pathak22 / ccnn Goto Github PK
View Code? Open in Web Editor NEW[ICCV 2015] Framework for optimizing CNNs with linear constraints for Semantic Segmentation
License: Other
[ICCV 2015] Framework for optimizing CNNs with linear constraints for Semantic Segmentation
License: Other
$ python ./src/train.py
cnn/ccnn/src/ccnn.py:19: RuntimeWarning: to-Python converter for std::vector<int, std::allocator > already registered; second conversion method ignored.
from python.ccnn import *
F0725 15:37:23.709703 4820 db.cpp:18] Check failed: status.ok() Failed to open leveldb /mnt/a/pathak/fcn_mil_cache/VOC2012/images_train_lmdb
IO error: /mnt/a/pathak/fcn_mil_cache/VOC2012/images_train_lmdb/LOCK: No such file or directory
*** Check failure stack trace: ***
@ 0x7f0e98f9e04d google::LogMessage::Fail()
@ 0x7f0e98fa0603 google::LogMessage::SendToLog()
@ 0x7f0e98f9dbdb google::LogMessage::Flush()
@ 0x7f0e98f9f54e google::LogMessageFatal::~LogMessageFatal()
@ 0x7f0e99320e23 caffe::db::LevelDB::Open()
@ 0x7f0e9937dfb6 caffe::DataLayer<>::DataLayerSetUp()
@ 0x7f0e9938e882 caffe::BaseDataLayer<>::LayerSetUp()
@ 0x7f0e9938e919 caffe::BasePrefetchingDataLayer<>::LayerSetUp()
@ 0x7f0e99367e99 caffe::Net<>::Init()
@ 0x7f0e9936a05e caffe::Net<>::Net()
@ 0x7f0e993000a5 caffe::Solver<>::InitTrainNet()
@ 0x7f0e9930138e caffe::Solver<>::Init()
@ 0x7f0e99301556 caffe::Solver<>::Solver()
@ 0x7f0e9992edf0 caffe::GetSolver<>()
@ 0x7f0e998fc7d3 caffe::GetSolverFromString()
@ 0x7f0e999086b8 boost::python::objects::caller_py_function_impl<>::operator()()
@ 0x7f0e987175cd boost::python::objects::function::call()
@ 0x7f0e987177c8 (unknown)
@ 0x7f0e9871f823 boost::python::detail::exception_handler::operator()()
@ 0x7f0e3a8a8d20 boost::python::detail::translate_exception<>::operator()()
@ 0x7f0e3a884f3d boost::_bi::list3<>::operator()<>()
@ 0x7f0e3a871ec7 boost::_bi::bind_t<>::operator()<>()
@ 0x7f0e3a86b268 boost::detail::function::function_obj_invoker2<>::invoke()
@ 0x7f0e9871f5dd boost::python::handle_exception_impl()
@ 0x7f0e98714999 (unknown)
@ 0x7f0e9b4567a3 PyObject_Call
@ 0x7f0e9b4ecb69 PyEval_EvalFrameEx
@ 0x7f0e9b4f24e9 PyEval_EvalCodeEx
@ 0x7f0e9b4f270a PyEval_EvalCode
@ 0x7f0e9b50b9cd run_mod
@ 0x7f0e9b50cb48 PyRun_FileExFlags
@ 0x7f0e9b50dd68 PyRun_SimpleFileExFlags
已放弃 (核心已转储)
请问这个问题什么意思?该怎么解决啊?着急,在线等!!!
Hi,
I am trying to re-generate the numbers reported in your ICCV paper from scratch. Everything works fine when I use the pre-trained models. But when I train the models myself they seem to classify all pixels to the background. The only files I generated myself are LMDBs using the provided code (/extras/generate_lmdb.py).
Can you please also tell your loss value at 35,000 iterations?
Here is the snapshot of last 100 iterations in my case:
I0601 04:28:17.000000 34992 solver.cpp:489] Iteration 34700, lr = 1e-06
I0601 04:28:33.000000 34992 solver.cpp:214] Iteration 34720, loss = 0.0137566
I0601 04:28:33.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0137567 (* 1 = 0.0137567 loss)
I0601 04:28:33.000000 34992 solver.cpp:489] Iteration 34720, lr = 1e-06
I0601 04:28:53.000000 34992 solver.cpp:214] Iteration 34740, loss = 0.0162213
I0601 04:28:53.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0162215 (* 1 = 0.0162215 loss)
I0601 04:28:53.000000 34992 solver.cpp:489] Iteration 34740, lr = 1e-06
I0601 04:29:11.000000 34992 solver.cpp:214] Iteration 34760, loss = 0.0857187
I0601 04:29:11.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0857188 (* 1 = 0.0857188 loss)
I0601 04:29:11.000000 34992 solver.cpp:489] Iteration 34760, lr = 1e-06
I0601 04:29:29.000000 34992 solver.cpp:214] Iteration 34780, loss = 0.365484
I0601 04:29:29.000000 34992 solver.cpp:229] Train net output #0: loss = 0.365484 (* 1 = 0.365484 loss)
I0601 04:29:29.000000 34992 solver.cpp:489] Iteration 34780, lr = 1e-06
I0601 04:29:47.000000 34992 solver.cpp:214] Iteration 34800, loss = 0.427323
I0601 04:29:47.000000 34992 solver.cpp:229] Train net output #0: loss = 0.427323 (* 1 = 0.427323 loss)
I0601 04:29:47.000000 34992 solver.cpp:489] Iteration 34800, lr = 1e-06
I0601 04:30:05.000000 34992 solver.cpp:214] Iteration 34820, loss = 0.272257
I0601 04:30:05.000000 34992 solver.cpp:229] Train net output #0: loss = 0.272258 (* 1 = 0.272258 loss)
I0601 04:30:05.000000 34992 solver.cpp:489] Iteration 34820, lr = 1e-06
I0601 04:30:23.000000 34992 solver.cpp:214] Iteration 34840, loss = 0.0660284
I0601 04:30:23.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0660286 (* 1 = 0.0660286 loss)
I0601 04:30:23.000000 34992 solver.cpp:489] Iteration 34840, lr = 1e-06
I0601 04:30:40.000000 34992 solver.cpp:214] Iteration 34860, loss = 0.696421
I0601 04:30:40.000000 34992 solver.cpp:229] Train net output #0: loss = 0.696421 (* 1 = 0.696421 loss)
I0601 04:30:40.000000 34992 solver.cpp:489] Iteration 34860, lr = 1e-06
I0601 04:30:58.000000 34992 solver.cpp:214] Iteration 34880, loss = 0.096779
I0601 04:30:58.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0967792 (* 1 = 0.0967792 loss)
I0601 04:30:58.000000 34992 solver.cpp:489] Iteration 34880, lr = 1e-06
I0601 04:31:16.000000 34992 solver.cpp:214] Iteration 34900, loss = 0.216992
I0601 04:31:16.000000 34992 solver.cpp:229] Train net output #0: loss = 0.216992 (* 1 = 0.216992 loss)
I0601 04:31:16.000000 34992 solver.cpp:489] Iteration 34900, lr = 1e-06
I0601 04:31:33.000000 34992 solver.cpp:214] Iteration 34920, loss = 0.047195
I0601 04:31:33.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0471952 (* 1 = 0.0471952 loss)
I0601 04:31:33.000000 34992 solver.cpp:489] Iteration 34920, lr = 1e-06
I0601 04:31:51.000000 34992 solver.cpp:214] Iteration 34940, loss = 0.32279
I0601 04:31:51.000000 34992 solver.cpp:229] Train net output #0: loss = 0.32279 (* 1 = 0.32279 loss)
I0601 04:31:51.000000 34992 solver.cpp:489] Iteration 34940, lr = 1e-06
I0601 04:32:08.000000 34992 solver.cpp:214] Iteration 34960, loss = 0.0153671
I0601 04:32:08.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0153672 (* 1 = 0.0153672 loss)
I0601 04:32:08.000000 34992 solver.cpp:489] Iteration 34960, lr = 1e-06
I0601 04:32:26.000000 34992 solver.cpp:214] Iteration 34980, loss = 0.0963567
I0601 04:32:26.000000 34992 solver.cpp:229] Train net output #0: loss = 0.0963568 (* 1 = 0.0963568 loss)
I0601 04:32:26.000000 34992 solver.cpp:489] Iteration 34980, lr = 1e-06
35000 iterations t = 892.211614132
Any help is welcome.
Thanks!
Hi,
I am trying to reproduce your work. During this period, sadly, I failed to load python.caffe in ccnn.py. It seems that this lib cannot be found. Besides, I am confused about how you implement the latent distribution optimisation.
Looking forward to your reply. Thank you~
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.