Skip to content

voxtral_tts: enable CUDA backend with 4w quantization (Ampere + Blackwell pre-exported artifacts) #3269

voxtral_tts: enable CUDA backend with 4w quantization (Ampere + Blackwell pre-exported artifacts)

voxtral_tts: enable CUDA backend with 4w quantization (Ampere + Blackwell pre-exported artifacts) #3269