Views
No views yet
mms-tts-* conversions (fp32 + q8 dynamic quantization).
Converted with the transformers.js 3.8.1 conversion script (torch 2.6.0
legacy exporter, onnxslim skipped, fp16 omitted -- see quantize_config.json).1import { pipeline } from '@huggingface/transformers';
2const synth = await pipeline('text-to-speech', 'payam1394/traxlate-mms-tts-ben', { dtype: 'q8' });