NLLB License

[sbmaruf]

❓ Questions and Help

Here is the NLLB model's license, https://github.com/facebookresearch/fairseq/blob/nllb/LICENSE.model.md

Can we use NLLB model output (translation from language X to language Y) to train a model and release that model under a Commercially permitted license (i.e., Apache 2.0)? I understand the model license is for Attribution-NonCommercial 4.0 International but to generate the model output, we are actually paying for the compute hours. I understand that we cannot use model weights for commercial purposes. But what about the output generated by the model?

[jav-ed]

I was looking forward to using NLLB, but then saw

All models are licensed under CC-BY-NC 4.0

If you have a look at https://www.gnu.org/licenses/gpl-faq.en.html#GPLOutput, you see

In general this is legally impossible; copyright law does not give you any say in the use of the output people make from their data using your program. If the user uses your program to enter or convert her own data, the copyright on the output belongs to her, not you. More generally, when a program translates its input into some other form, the copyright status of the output inherits that of the input it was generated from.

Still, I would appreciate it, if the developer or Meta would allow us to use their Model under MIT or to pursue the idea suggested by @sbmaruf .

An answer to that question would be nice - thank you.

[jav-ed]

Just to have it mentioned once, even though I am aware of https://www.gnu.org/licenses/gpl-faq.en.html#GPLOutput, I still do not feel comfortable using the models. Obviously, the intention of Meta or the developer was to hinder people from using it for commercial purposes. Thus, I would like to follow CC-BY-NC 4.0 and not use it for commercial purposes.

So for me and for a lot of other people, it would be great to have the models under MIT or Apache.