Adds radio data recipe #1400
Adds radio data recipe #1400
Conversation
lhotse/recipes/radio.py
Outdated
| from lhotse.supervision import SupervisionSegment, SupervisionSet | ||
| from lhotse.utils import Pathlike | ||
|
|
||
| set_ffmpeg_torchaudio_info_enabled(False) |
There was a problem hiding this comment.
One last request - don’t set these things in the global scope because it’ll be executed when Lhotse is imported. Move it to local scope (even better i think this one works like a context manager)
There was a problem hiding this comment.
I think I fixed it now?
There was a problem hiding this comment.
Almost - now you need to move to actual usage to inside the function that does the audio reading and use it with Python with statement
There was a problem hiding this comment.
I actually forget why I started adding this in to all the Lhotse recipes. I think it fixed some problem I was having at some point, but I have no recollection of what that problem was. Do I actually need it? When I actually use it locally each time it gives me a ton of horrible logging messages.
There was a problem hiding this comment.
You shouldn’t need it. We are using libsndfile now by default rather than torchaudio. Feel free to remove it if you prefer
There was a problem hiding this comment.
That is what I ended up doing. I believe the current updated commit just removed this.
…enabled call. The recipe runs fine without it.
|
Is there anything left to do for this one? |
|
Looks good, thanks! |
This is the recipe for some radio data we collected and are able to distribute if someone emails us for it. We may possibly have a way of releasing it with some restrictions in the future, but for now it is just a prepare script. The paper it corresponds to is
https://aclanthology.org/2024.naacl-long.286.pdf
The main use is for geolocating speech, self-supervised model pretraining, or language id.