KeyError:"Configuration file contains the following unrecognized options: ['log']

[Anker-Lee]

Hello!
When I run the 7th command: tune_rst_parser rst_discourse_tb_edus_TRAINING_TRAIN.json rst_discourse_tb_edus_TRAINING_DEV.json rst_parsing_model,
it gives me theKeyError. Anybody can help me to solve this problem?

[desilinguist]

Ah, this is because rstfinder hasn't been updated to use the new release of the SKLL that recently deprecated some fields.

I'll work on a fix for this. In the meantime, you can simply downgrade SKLL to v2.5 in your conda environment and things should work again.

[Anker-Lee]

Thank you very much!
Downgrade SKLL to v2.5 works quite well!
But when I ran the 7th command,
tune_rst_parser rst_discourse_tb_edus_TRAINING_TRAIN.json rst_discourse_tb_edus_TRAINING_DEV.json rst_parsing_model
It ran well at the beginning, but when it came to around 54.872392%, it stopped for a short moment and threw out concurrent.futures.proess.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending error.
Could you please help me out again?

[desilinguist]

Very strange. I've never seen that error before. Are there any log files created perhaps? If so, please share.

[Anker-Lee]

Traceback (most recent call last): File "/home/anker/anaconda3/envs/rstenv2/bin/tune_rst_parser", line 11, in <module> sys.exit(main()) File "/home/anker/anaconda3/envs/rstenv2/lib/python3.8/site-packages/rstfinder/tune_rst_parser.py", line 306, in main for C_value, results in zip(C_values, all_results): File "/home/anker/anaconda3/envs/rstenv2/lib/python3.8/concurrent/futures/process.py", line 484, in _chain_from_iterable_of_lists for element in iterable: File "/home/anker/anaconda3/envs/rstenv2/lib/python3.8/concurrent/futures/_base.py", line 619, in result_iterator yield fs.pop().result() File "/home/anker/anaconda3/envs/rstenv2/lib/python3.8/concurrent/futures/_base.py", line 437, in result return self.__get_result() File "/home/anker/anaconda3/envs/rstenv2/lib/python3.8/concurrent/futures/_base.py", line 389, in __get_result raise self._exception concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.

[Anker-Lee]

Very strange. I've never seen that error before. Are there any log files created perhaps? If so, please share.

It seems that there was not log files created. Is it possible that it can not work on a virtual machine?

[desilinguist]

Oh interesting. I’ve never tried it on a VM. That could be it, I’m not sure. Is the error reproducible - if you run it again, does it fail again with the same error?

…

________________________________ From: Ankery ***@***.***> Sent: Sunday, February 20, 2022 12:55:16 AM To: EducationalTestingService/rstfinder ***@***.***> Cc: Nitin Madnani ***@***.***>; Comment ***@***.***> Subject: Re: [EducationalTestingService/rstfinder] KeyError:"Configuration file contains the following unrecognized options: ['log'] (Issue #73) Very strange. I've never seen that error before. Are there any log files created perhaps? If so, please share. It seems that there was not log files created. Is it possible that it can not work on a virtual machine? — Reply to this email directly, view it on GitHub<#73 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AABHPEDUOGIL6WWOWMYQOO3U4B64JANCNFSM5OX5VHQA>. You are receiving this because you commented.Message ID: ***@***.***>

[Anker-Lee]

Oh interesting. I’ve never tried it on a VM. That could be it, I’m not sure. Is the error reproducible - if you run it again, does it fail again with the same error?
…

Yes, it is reproducible.

[desilinguist]

Could you try on a regular machine if possible and see if you get the same error?

I'll also try to run it myself.

[Anker-Lee]

Could you try on a regular machine if possible and see if you get the same error?

I'll also try to run it myself.

It seems that memory distributed to the VM was too small, I have redistributed the memory. I will try it again. THANKS A LOT FOR HELPING ME! : )

[Anker-Lee]

Could you try on a regular machine if possible and see if you get the same error?

I'll also try to run it myself.

It does not throw errors at that position!
Is it normal for the training and featuring process to wait for a while?
It stopped at INFO: Hyperparameters: C:0.25 ... warm_start: False for a long time.
It seems that the training process takes quite a long time.

[Anker-Lee]

Could you try on a regular machine if possible and see if you get the same error?

I'll also try to run it myself.

Thanks a lot for your helpful tool!
After tune_rst_parser rst_discourse_tb_edus_TRAINING_TRAIN.json rst_discourse_tb_edus_TRAINING_DEV.json rst_parsing_model , it throw out
ValueError: X has 640 features, but StanderScaler is expecting 74698 features as input.
I seen the answer in #72, is skill v2.5 still can not work?

[desilinguist]

Yeah, looks like there’s some incompatibility. I am working on fixing this now.

[Anker-Lee]

Thanks a lot for helping me!
You are really nice!

[desilinguist]

You're most welcome!