Closed kwonej0617 closed 1 year ago
Hi @kwonej0617, this is the first time I encounter this segmentation fault issue. Can you show me the first couple of lines of data.json, data.readcount, and data.index files?
Hi @chrishendra93! Thank you for your quick reply.
They look fine to me. Can you run it without the --infer_mod-rate option?
I got the same Segmentation fault even in running without --infer_mod-rate option. Actually, the Segmentation fault appears when I run the command, m6anet-run_inference without any requirement and option.
Do you know how much memory it takes when running m6anet-run_inference?
I am not too sure, should not be that much. May I know your RAM size? Can you try reducing the batchsize? This can be done with --batch_size option, maybe try --batch_size 1 or some other smaller number?
Hi,
Is this a proper way to check RAM size? [ek81w@ghpcc06 m6anet]$ grep MemTotal /proc/meminfo MemTotal: 264,498,876 kB
Also, I run the m6anet-run_inference --input_dir dataprep_out --out_dir inference_out --infer_mod-rate --n_processes 4 --batch_size 1 and also got Segmentation fault error. :(
This feels like a cpp related error instead of python especially since batch size 1 does not work. Can you show me the full error, also do you happen to have another machine that you can try for this?
Conda is an open source package management system and environment management system.
To use conda packages, first run source /share/pkg/condas/2018-05-11/bin/activate
Then, to list available environments conda env list
To activate a conda enviroment source activate [environment_name] e.g.: source activate biopython_1.71_py35h637b7d7_0
Your current environment will be displayed on your shell prompt.
To configure conda so you can install your own environments, you can run:
cp /share/pkg/condas/2018-05-11/condarc_example ~/.condarc
This config tells conda to install environment and package files under a .conda directory in your home, and allow you to install your own environments. Modify the .condarc file to change where your conda environment files are stored.
Also, my colleague run the command line for me and she got the same error message.
Okay let me check this on my machine. Are you running the demo data? Which os are you using?
Yes, I was running the demo data provided. I used linux.
Thank you so much!
Sorry can you also tell me which python version are you using?
Thank you!
No problem. I used python=3.8.
Thank you so much for your help!
Hello, Christopher!
I just wanted to check if you have solved this problem. I asked IT administration in my school about Segmentation fault issue. He said it's caused by a glibc mis-match and recommend that I should use a singularity container with the appropriate package versions. Do you think it would resolve the problem? Also, do you have a plan to make a singularity container?
I am looking forward to hearing from you! Thank you.
Hi @kwonej0617, sorry but currently I am overseas so it is kind of hard to troubleshoot this. I am going to be back next week and I can take a look after. There was a plan to make a docker container for this too
No problem! I am looking forward to hearing from you! Have a safe trip.
hi @kwonej0617, sorry for the delay
I could not replicate the error message with linux, but just in case, I have updated the package because I realized there was a missing dependency in setup.py file. Can you try installing the latest version through pip / from running python setup.py install on a new conda environment to see if it's working?
Thank you
Regards
Christopher Hendra
hi @kwonej0617 , from your recent issues, I am assuming that you have managed to run m6anet on your machine? If so I am going to close this issue
Thanks!
Hi Developer!
I was running m6anet with demo data.
m6anet-dataprep appeared to run successfully though it had a warning message. m6anet-dataprep --eventalign demo/eventalign.txt --out_dir dataprep_out --n_processes 4
/home/ek81w/.conda/envs/m6anet/lib/python3.8/site-packages/m6anet/scripts/dataprep.py:143: SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy chunk_split['line_length'] = np.array(lines) /home/ek81w/.conda/envs/m6anet/lib/python3.8/site-packages/m6anet/scripts/dataprep.py:101: PerformanceWarning: indexing past lexsort depth may impact performance. pos_end += eventalign_result.loc[index]['line_length'].sum()
However, when I run m6anet-run_inference, it showed Segmentation fault message. m6anet-run_inference --input_dir dataprep_out --out_dir inference_out --infer_mod-rate --n_processes 4 I tried increasing the memory request, but it didn't work. By chance, if you know an idea to fix this problem, please let me know.
Thank you.
EuiJin