Streamalign-SLM-R16 / README.md
dd3434's picture
Upload README.md with huggingface_hub
357f31e verified
|
Raw
History Blame
405 Bytes

Streamalign-SLM (R16)

StreamAlign speech language model (Llama-3.2-1B backbone, R16 hier durfirst-durreg, text-KL recipe). Scores speech via the paired R16 RVQ teacher Streamalign-R16.

  • Original checkpoint: R16_step180k_lossWtext5_kl05_210k/step_00185000.pt
  • Speech tokenizer / teacher: Streamalign-R16
  • Reported metrics: SALMon 69.1 (4-task, no pitch), StoryCloze 72.1 (sSC 59.0 / tSC 85.1)