Comma when using lm_optimize?

[FrontierDK]

Hi all :)

Is there a limitation on what can be used in the sample.csv file? I am trying to use strings with a comma, say something like this:

wav_filename,wav_filesize,transcript
files/yes_this_is_a_test.wav,94140,yes, this is a test

Also, I add the exact same line to language_model.txt

I can run both the generate_lm.py and generate_scorer_package - but when it gets to lm_optimize - I get this error:
TypeError: sequence item 2: expected str instance, float found

How does one use a comma in the text to be learned from? If I remove the comma, all runs well.

[reuben]

Add quotes around the transcript:

 files/yes_this_is_a_test.wav,94140,"yes, this is a test"

[FrontierDK]

Thank you VERY much 👍