Apache-2.0
Google / MADLAD-400 authors; optimized ONNX export and INT4/INT8 quantization by Luca (niedev/OnnxModelsEnhancer). Downloaded directly from the exporter’s v1.0.0-beta release. The five model files are byte-for-byte identical to the previously used mirror, as verified by SHA-256. No changes to model weights by JWH Software.
These third-party rights are independent of purchasing SpeakMyLanguage. No endorsement is implied.
ec3183e6106182938fe095d8e48057e2412ded278b560c2dc818c72cfedd682d