Licensing Clarification
I just wanted to ask for some clarification around licensing and commercial use, as there seem to be a few mixed messages.
I noticed that the terms say:
“Researcher shall use the Database only for non-commercial research and educational purposes.”
But elsewhere it’s mentioned:
“Speech recognition models trained on the GigaSpeech dataset may be eligible for commercial license, provided they abide by the 'Fair Use' terms...”
Also, since the dataset is on HuggingFace under an Apache 2.0 licence (which I believe allows commercial use), I’m a bit confused about what’s actually permitted.
Could you clarify:
- Is it okay to use GigaSpeech to train models for commercial use?
- Does the Apache 2.0 licence cover everything in the dataset, including audio files?
- Are there any extra restrictions we should keep in mind apart from general copyright rules?
Would really appreciate any guidance on this. Thanks in advance!
I was also wondering about that.
The HF dataset here specifies the Apache 2 licence. This is specifically about the data itself, not just the code. But this might just be a mistake? Because the authors likely don't have the copyright on the data to redeclare it under Apache 2.
Apache 2 specifically allows also commercial use. So that is in contradiction to the terms of access. What takes precedence here?
Apache 2 also allows to redistribute the data, without such terms of access that are given here. But I guess this is not allowed?
So basically, it is not correct that the data is under Apache 2. This should maybe be fixed? This line here:
https://huggingface.co/datasets/speechcolab/gigaspeech/blob/main/gigaspeech.py#L84