Inquiry About AstraMindAI/xtts2-gpt Model Training and Licensing
hope this message finds you well. I am interested in using the AstraMindAI/xtts2-gpt model, available on Hugging Face, for a commercial text-to-speech application. The model’s capabilities, particularly its voice cloning and multilingual support, are impressive and align well with my project requirements.
To ensure compliance with licensing terms, I have a few questions regarding the model’s training process and data:
Does the xtts2-gpt model rely on Coqui’s pre-trained XTTS-v2 weights (subject to the Coqui Public Model License) for training or fine-tuning?
If not, was the model trained from scratch or fine-tuned using datasets that explicitly permit commercial use (e.g., LibriSpeech under CC BY 4.0 or Common Voice under CC0)?
Are there any additional licensing considerations or restrictions I should be aware of when using xtts2-gpt for commercial purposes?
Your clarification will help ensure that my use of the model complies with all relevant licenses and minimizes any legal risks. If possible, please confirm whether the model was trained independently of Coqui’s restricted weights and whether the training datasets are compliant with commercial use.
Thank you for your time and assistance. I look forward to your response and am happy to provide more details about my use case if needed. You can reach me at [Your Email Address] or reply directly to this message.