---
title: nemotron-3.5-asr-streaming-multilingual-0.6b-coreml
canonical_url: "https://www.modelscope.cn/models/Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml"
md_url: "https://www.modelscope.cn/models/Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml.md"
repository: Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml
chinese_name: "Nemotron 3.5 ASR Streaming Multilingual 0.6B CoreML"
last_updated: 2026-08-26
license: other
pipeline_tag: auto-speech-recognition
tasks:
  - auto-speech-recognition
downloads: 4
stars: 0
tags:
  - fluidaudio
  - coreml
  - apple-silicon
  - ane
  - streaming-asr
  - rnnt
  - multilingual
---

# nemotron-3.5-asr-streaming-multilingual-0.6b-coreml

> nemotron-3.5-asr-streaming-multilingual-0.6b-coreml - Lorqa 在 ModelScope 开源的模型。NVIDIA Nemotron 3.5 Streaming Multilingual 0.6B 的 CoreML 导出（2240ms 流式档）。供 Lorqa 在 Apple 平台（macOS / iOS）做端侧实时语音识别使用，通过 FluidAudio SDK 在 Apple Neural Engine 上运行。支持 40+…

Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml 是 ModelScope 魔搭社区上的auto-speech-recognition模型，采用 other 许可。

- **Repository**: Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml
- **License**: other
- **Tasks**: auto-speech-recognition
- **Tags**: fluidaudio, coreml, apple-silicon, ane, streaming-asr, rnnt, multilingual
- **Downloads**: 4
- **Stars**: 0
- **Last updated**: 2026-08-26

Source: https://www.modelscope.cn/models/Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml

---

# Nemotron 3.5 ASR Streaming Multilingual 0.6B — CoreML (2240 ms tier)

CoreML build of NVIDIA's Nemotron 3.5 streaming ASR model, used by
[Lorqa](https://lorqa.app) for on-device live captions and audio transcripts on
macOS and iOS. Runs on the Apple Neural Engine through the
[FluidAudio](https://github.com/FluidInference/FluidAudio) SDK.

- 40+ languages with automatic language detection (no per-language download)
- Cache-aware streaming encoder: state is carried across chunks, so decoding is
  incremental rather than re-running on a growing buffer
- Fully on-device — audio never leaves the machine

## What's in here

Only the **`multilingual/2240 ms`** tier — 22 files, ~634 MB.

```
multilingual/2240ms/
├── encoder.mlmodelc/          # 539 MB
├── decoder.mlmodelc/
├── decoder_joint.mlmodelc/
├── joint.mlmodelc/
├── preprocessor.mlmodelc/
├── metadata.json              # prompt_dictionary, lang tag token ids
└── tokenizer.json
```

Upstream ships 8 bundles (`latin` / `multilingual` × 560/1120/2240/4480 ms).
The other seven are deliberately not mirrored here — see below.

## Why the 2240 ms tier

The four tiers are not quality levels or versions. They are **streaming chunk
lengths**: how much audio goes into one forward pass (`chunk_mel_frames` ×
10 ms hop). All four share the same `att_context_size` of `[42, 13]`.

Measured on the multilingual bundle (RTFx = audio-seconds per wall-second,
higher is better; parenthesised figure is WER, or CER for zh/ja — lower is
better):

| Tier | Chinese | Japanese | Multilingual-English |
|---|---|---|---|
| 560 ms | 22 (19.48%) | 21 (14.61%) | 23 (9.15%) |
| 1120 ms | 27 (18.75%) | 26 (13.77%) | 71 (8.64%) |
| **2240 ms** ⭐ | **89 (18.57%)** | **84 (13.79%)** | **80 (8.76%)** |
| 4480 ms | 90 (18.05%) | 89 (13.82%) | 78 (8.78%) |

2240 ms is the upstream-recommended default and the only tier that is strong on
both axes for CJK. 560 ms falls off the trained attention tiling, so it is worse
on accuracy *and* throughput — it is not a lower-latency trade, it is dominated.
Figures are from the upstream model card, not re-measured here.

## Provenance

NVIDIA `nemotron-3.5-asr-streaming-0.6b` → CoreML export by
[FluidInference](https://huggingface.co/FluidInference) → this mirror. Weights
are unmodified; this repo exists so Lorqa pins a stable source and users in
mainland China get a fast one. The same content is mirrored on
[HuggingFace](https://huggingface.co/Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml)
and [ModelScope](https://www.modelscope.cn/models/Lorqa/nemotron-3.5-asr-streaming-multilingual-0.6b-coreml).

Licensed under OpenMDW 1.1, following upstream.