                 Changes in version 0.1.0 (2026-10-14)                  

  - First CRAN release.

                       Changes in version 0.0.1.1                       

  - On CUDA, encoder layers run as one traced TorchScript function
    shared by all 31 layers, and each chunk ends with a minor garbage
    collection. Peak device memory drops from 1.7-2.1 GiB to 0.47-0.55
    GiB (weights are 0.37 GiB), and a streaming step from 55 ms to
    about 18 ms on an RTX 5060 Ti. load_n3d() traces and warms up the
    layer once (about 1 s). options(n3d.jit = FALSE) runs the plain R
    functions; the CPU always does.
  - Log-mel features and logits for the whole input stay on the CPU, and
    one chunk at a time moves to the device, so device memory no longer
    grows with the length of the audio (0.52 GiB for 9.8 min).

                        Changes in version 0.0.1                        

  - Initial version: native R 'torch' port of NVIDIA Nemotron 3
    Diarization with offline diarization (diarize()), streaming sessions
    (n3d_stream(), n3d_stream_push(), n3d_stream_finish()), log-mel
    features and the arrival-order speaker cache implemented in R, and
    pinned-revision weight download through 'hfhub'.
