Closed bhargava-morampalli closed 2 years ago
hi @bhargava-morampalli , can I ask if you modify some of the codes because I think the script name should be m6anet-run_inference? Anyway it seems that you are passing None type to the os.path.join argument somewhere in the script, is it possible that the problem is because you use -i and -o instead of --input_dir and --output_dir, that's why the command line does not recognize the input and output directory argument?
Oh, sorry - I just followed the quick start thing. So, I should use run_inference instead of inference?
m6anet-inference -input_dir demo_data --out_dir demo_data ---n_processes 4
I have not touched any code. Also, for inference - --input_dir and --out_dir did not work - that's why I changed them to -i and -o. I will try with run_inference and let you know if it works.
Okay, I used the m6anet-run_inference and I got the data.result.csv.gz file. It's very small and only a few bytes. Is this normal? also, n_processes is for allocating the threads - is it correct?
Ah I see, it's weird that it can still execute m6anet-inference, perhaps I have forgotten to clear some files / cache, let me check that on my end. Apology for the typo in the documentation, I have just updated it so that people will not mistake the command. Thanks!
Anyway, you are right, --n_processes is to allocate the number of threads. Also, by default m6anet will require each position to have at least 20 reads. May I know the size of the data.readcount files (like how many rows and whether they seem to have a lot of positions with at least 20 reads)? Also, can you check if the entries inside data.result.csv.gz make sense?
That's great, thanks. I will check the data.readcount and also about the data in csv.gz. I will let you know the results.
oh right, please try out with --input_dir instead of -input_dir as stated in the documentation before, that was a typo that I have corrected in the documentation, again, sorry for this
hi @bhargava-morampalli , can I ask if you have managed to run this successfully? If you have, then I want to close this issue, otherwise please let me know of any problems you are facing with running m6anet
Hi Chris,
Thanks for releasing the new model for m6anet. I have run the nanopolish, data-prep steps successfully but I am encountering error with the inference step. Here's the command I used and the error that I got.
Here are the contents of the m6anet data-prep output directory.
and a few lines at the beginning for each file. data.index
data.json
data.log
data.readcount
eventalign.index
Happy to provide more details if needed. Please let me know how to resolve this error.