vidit09 / geoshift Goto Github PK
View Code? Open in Web Editor NEWGeoShift CVPR-2023
GeoShift CVPR-2023
hello,i want to aks a question.
In the last trainging -meaning teacher training,i wonder if IMS_ PER_ BATCH: Can 2 be changed to 1?
Looking forward to your reply。
python train_net.py --config-file configs/base_city_rcnn_person_random_crop.yaml
Traceback (most recent call last):
File "train_net.py", line 13, in
from data.datasets import builtin
ModuleNotFoundError: No module named 'data.datasets'
i wonder how can i solve such question?
How can I obtain an image similar to the one labeled as "ours" in Figure 5 of your paper?
FloatingPointError: Predicted boxes or scores contain Inf/NaN. Training has diverged.
[04/15 14:33:06 d2.engine .hooks]: Overall training speed: 2 iterations in 0:00:07(3.7250 s / it)[04/15 14:33:06 d2.engine.hooks]: Total training time: 0:00:07 (0:00:00 on hooks)
[04/15 14:33065 62.tils.events]: eta: 1 day,0:25:08 iter: 4 total_ loss: 1.289e106 loss cls: 1.162e+05 loss_box_ reg: 1.235e+04loss_rpn_cls: 3.925e+C5 loss-rp_loc: 7.65e+C5 time: 3.636 last_time:0.2101 data_time: 4.9687 last data _time: 0.0o4 lr: 3.997e-07max_mem: 4413M
Traceback (most recent call last):
File "D: \geoshift-main\geoshift-main\train_net.py", line 55,in
launch(
File "d: \temp\detectron2-main\detectron2-main\detectron2\engine\launch.py", line 84,in launchmain_func(*args)
File "D: \geoshift-main\geoshift-main\train_net.py", line 49,in maintrainer.train()
File "d:\temp\detectron2-main\detectron2-main\detectron2\engine\defaults.py", line 486,in trainsuper().train(self.start_iter, self.max_iter)
File "d : \temp\detectron2-main\detectron2-main(detectron2\engine\train_loop.py", line 155,in trainself.run_step(
File "d: \temp\ldetectron2-main(detectron2-main(detectron2\engine\defaults.py", line 496,in run_stepself._trainer.run_step()
File "d: \temp\detectron2-main\detectron2-main(detectron2\engine\train_loop.py" , line 310,in run_steploss_dict = self.model(data)
File "D: \anaconda\envs\geo\lib\site-packages\torch \n(modules[module.py" , line 1102,in _call_implreturn forward_call(*input,**kwargs)
File "d:\temp(detectron2-main\detectron2-main)detectron2\modeling(meta_arch\rcnn.py""line 161,in forwardproposals, proposal_losses = self. proposal_generator(images,features, gt_instances)
File"D: \anaconda\envs\geo\lib\site-packages \torch\nn\modules\module.py", line 1102,in _call_implreturn forward_call(*input,**kwargs)
File "d:\temp(detectron2-main(detectron2-main\detectron2[modeling\proposal generator\rpn.py" line 477,in forwardproposals = self.predict_proposals(
File "d: templdetectron2-main\detectron2-main\detectron2[modelingIproposal generatorIrpn.py", line 503,in predict_proposalsreturn find_top_rpn_proposals(
File "d: temp]detectron2-main(detectron2-main detectron2[modelinglproposal generator \proposal_utils.py", line 108,in find_top_rpn_proposals
raise FloatingPointError(
FloatingPointError: Predicted boxes or scores contain Inf/NaN. Training has diverged.
when i try python train_net.py --config-file configs/base_city_rcnn_person_random_crop.yaml,i meet such error,i wonder how can i solve it.
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.