Common voice dataset 15.0 version release

[anz2]

Hi All and especially @reach-vb!
I need some help with the common voice dataset. The Georgian dataset has increased recently up to 154 hr total (Common Voice). I am exploring the options for STT model training. I have seen the 13.0 version of the common voice dataset on huggingface and I was wondering if 15.0 is in progress and will see it soon.

I also see the option for loading locally similar to this thread:

Unable to load CommonVoice latest version

This line of code works well with commonvoice version 6.1.0 but i get error while using for version .70.0 from datasets import load_dataset, load_metric, Audio common_voice_train = load_dataset("common_voice", "lg", split="train+validation",version="7.0.0") common_voice_test = load_dataset("common_voice", "lg", split="test",version="7.0.0") ValueError: Cannot name a custom BuilderConfig the same as an available BuilderConfig. Change the name. Available BuilderConfigs: ['ab', 'ar', 'as', 'br'…

But I am wondering if the latest dataset version will be added soon.
I would try to do PR but the 15.0 dataset repo needs to be created first as for 13.0 (mozilla-foundation/common_voice_13_0).

Thanks!

[mariosasko]

@reach-vb should be able to help with that.