Closed smeylan closed 4 years ago
Hi, Leandro here from ROS lab.
Having the exact same output.
This is tmp_data tree before it gets deleted
tmp_data/ ├── features │ ├── ALUCs_out_individual.txt │ └── final_feats.txt ├── helenab-a1-nsm-chops013.wav ├── helenab-a1-nsm-chops014.wav └── short ├── helenab-a1-nsm-chops013_00344150_00358610.wav ├── helenab-a1-nsm-chops013_00374930_00418520.wav ├── helenab-a1-nsm-chops013_00439980_00450110.wav ├── helenab-a1-nsm-chops014_00129920_00145090.wav ├── helenab-a1-nsm-chops014_00224930_00229920.wav ├── helenab-a1-nsm-chops014_00240010_00243250.wav ├── helenab-a1-nsm-chops014_00244300_00245550.wav ├── helenab-a1-nsm-chops014_00246390_00248480.wav ├── helenab-a1-nsm-chops014_00250100_00254710.wav ├── helenab-a1-nsm-chops014_00661250_00662460.wav ├── helenab-a1-nsm-chops014_00662530_00664960.wav ├── helenab-a1-nsm-chops014_00665110_00668370.wav ├── helenab-a1-nsm-chops014_00669670_00682140.wav ├── helenab-a1-nsm-chops014_00690590_00719260.wav ├── helenab-a1-nsm-chops014_00719310_00729940.wav ├── helenab-a1-nsm-chops014_00801410_00810200.wav ├── helenab-a1-nsm-chops014_00823250_00854840.wav ├── helenab-a1-nsm-chops014_00983510_00984590.wav ├── helenab-a1-nsm-chops014_00984930_00990110.wav ├── helenab-a1-nsm-chops014_01190030_01191630.wav └── helenab-a1-nsm-chops014_01194900_01199910.wav
Hi! Thanks for testing this out! Could you paste contents of sylnet.log here after the crash? Thanks!
I found the bug.
Line 62, when running SylNet you are using &> to save stdout and stderr to file.
This is fine using bash but not using sh (it is not the same thing :D!). In sh this will run SylNet in background and continue processing main script which end up in error since SylNet_out_files.txt has not been created yet.
In your README below Usage you run the commands using sh when you should be running them using bash (as the shebang you defined).
I would make run_alice.sh executable and run it like ./run_alice.sh (since the shebang is well defined) . Not sure if this works properly in Mac though.
Thanks for the update! Let me look into this. I need to figure out how the sh/bash differ in practice, so this might take a while. (me being a Mac user and stuff)
No problem, I will test ALICE with some short unseen anotated data and update you with the results. I made a quick test of the diarization tool and it had promising results as well.
Indeed chmod +x run_ALICE.sh
and ./run_ALICE.sh demo/ROS_5271_20_01_03600_snippet_mono.wav
fixes my problem -- thanks @macramole!
I see your point now! Thank you!
As a fix, I updated the readme and changed the permissions of the run_ALICE.sh for execution. I will have to study the shebang definitions if it is the correct one now or not, but at least it should work.
Cool. I'm pretty sure it is correct.
Running the demo (on Ubuntu 14.04),