self-ensemble-visual-domain-adapt
self-ensemble-visual-domain-adapt copied to clipboard
Britefury
Error for --exp=syndigits_svhn and --exp=synsigns_gtsrb
I got the following errors for the two experiments above.
".../domain_datasets.py", line 24, in get_config
raise ValueError('Could not find configuration file domain_datasets.cfg')
These are the two lines from batch_search_exp.sh that I tried running.
python experiment_domainadapt_meanteacher.py --exp=syndigits_svhn --log_file=results_exp_ctaa/log_syndigits_svhn_ctaa_run${2}.txt --standardise_samples --batch_size=128 --cls_balance=0.0 --num_epochs=300
python experiment_domainadapt_meanteacher.py --exp=synsigns_gtsrb --log_file=results_exp_ctaa/log_synsigns_gtsrb_ctaa_run${2}.txt --standardise_samples --batch_size=128 --cls_balance=0.0 --num_epochs=300
Could you help check why this might be the case?
Woops.... Forgot to put that in the README!
You will need to create a file called 'domain_datasets.cfg'.
In it put:
[paths]
syndigits=<syndigits path>
synsigns=<synsigns path>
gtsrb=<gtsrb path>
Thank you for your response!!
I looked into the code and believe that prepare_gtsrb.py and prepare_synsigns.py should be run to prepare as well (and I tried running them). By the way, is there prepare_syndigits.py to prep for syn_digits dataset?
I created domain_datasets.cfg file with the following content.
[paths]
syn_digits=data/syn_digits
syn_signs=data/syn_signs
gtsrb=data/gtsrb
Then I ran python prepare_synsigns.py. I got an error saying:
Labels path data/syn_signs/synthetic_data/train_labelling.txt does not exist
For prepare_gtsrb.py, I got
Creating data/gtsrb/gtsrb.h5...
ERROR!!! Training images path data/gtsrb/Final_Training/Images does not exist
Closing remaining open files:data/gtsrb/gtsrb.h5...done
As of now, when I run the two experiments mentioned previously I got the following errors. (data not found)
RuntimeError: Could not load Syn-Digits dataset
RuntimeError: Could not load Syn-Signs from data/syn_signs/syn_signs.h5; please run 'prepare_synsigns.py' to create it
Could you help update the README.md to help with instructions on how to get these datasets prepared? Thank you so much!!
I've updated the README to include the relevant info. I have also added a description of the directory layout that the software will expect. I hope this helps. Please let me know if not.
Oh, please delete data/gtsrb/gtsrb.h5 as it will likely be incomplete.
Thank you very much for the instructions. They're very helpful!
I was wondering it would be possible to get the actual dataset on synsigns from you as well since there's no public link. Please let me know! :)
Congratulations for your ICLR2018 acceptance! I also would really appreciate it if you could share the synsigns dataset with me.
Here:
https://drive.google.com/file/d/0B-N5tVpsXW5mb1l0SjJhMFJZVk0/view?usp=sharing