Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-1120ms-onnx-int8

Lorqa Nemotron 3.5 multilingual — 1120 ms ONNX INT8

Byte-identical runtime mirror of csukuangfj2/sherpa-onnx-nemotron-3.5-asr-streaming-0.6b-1120ms-int8-2026-06-11 at revision cba1c96ca5ef0e8393b50584ae153a79145dc492. The source model is NVIDIA Nemotron 3.5 ASR, licensed under OpenMDW 1.1. Original exporter attribution is retained in UPSTREAM_README.md.

The four runtime files total 682,215,474 bytes (about 682 MB). Encoder, decoder, joiner and tokens are unchanged. No training, re-quantization or accuracy improvements are claimed. Upstream sample recordings are not included in this mirror.

This is the multilingual model, including Mandarin Chinese; it is not the older English-only Nemotron. Chinese is a broad-coverage locale and accuracy needs evaluation on your recordings. Chunk size is 1120 ms, not a guarantee of end-to-end visible text latency. Larger chunks trade responsiveness for accuracy and compute efficiency; running memory exceeds download size.

Runtime compatibility

Use a sherpa-onnx release supporting Nemotron 3.5 multilingual, with encoder.int8.onnx, decoder.int8.onnx, joiner.int8.onnx and tokens.txt, and a 128-bin feature extractor. Per-stream language strings support explicit languages or auto. See the official multilingual deployment instructions.

This is an ONNX bundle, not a Core ML mlmodelc bundle; it cannot be loaded by Core ML or FluidAudio loaders.

Mirrors: Hugging Face · ModelScope. Verify SHA256SUMS after download. PROVENANCE.json records the upstream revision and intended chunk size.