v1.4.2

TimoBolkart/FLAME-Universev1.4.2Aug 14, 2026by LauraGPT

AI Summary

This release focuses on stability improvements for DeepSpeed training and ASR output alignment, introduces SRT subtitle support for native runtimes, and refreshes platform-specific runtime builds.

Key Highlights

Fixed split punctuation-token alignment in AutoModel
Corrected DDP gradient accumulation using no_sync
Added SRT subtitle output to native llama.cpp runtimes
Refreshed ggml/Vulkan baseline for AMD Windows stability
Validated native Fun-ASR-Nano serving on vLLM 0.27.1

New Features

SRT subtitle output support in command-line runtimes
Native Fun-ASR-Nano serving validated on vLLM 0.27.1
Product deployment paths for SenseVoice via audio.cpp and native server
Prebuilt llama.cpp runtimes for 9 platforms (Linux, macOS, Windows)

Full Release Notes

## FunASR 1.4.2

Install or upgrade the Python package:

```bash
pip install -U funasr==1.4.2
```

PyPI: https://pypi.org/project/funasr/1.4.2/

### Python package

- Fixes split punctuation-token alignment in `AutoModel` so punctuation results remain aligned with ASR output.
- Corrects DDP gradient accumulation by applying `no_sync` only to non-update microsteps.
- Initializes DeepSpeed mode from the resolved training configuration instead of relying on stale CLI state.

### Native and industrial deployment

- Adds SRT subtitle output to the native FunASR llama.cpp/GGUF command-line runtimes.
- Refreshes the ggml/Vulkan baseline to fix AMD Windows submission crashes reported by runtime users.
- Documents and validates native Fun-ASR-Nano serving on vLLM 0.27.1, while retaining the official-weight split-engine path.
- Adds product deployment paths for SenseVoice through audio.cpp and the native server, and clarifies the current speaker-diarization boundary.

### Prebuilt llama.cpp runtime

The release includes verified `runtime-llamacpp-v0.2.0` builds for nine Linux, macOS, and Windows targets, including Windows CUDA and Linux/Windows Vulkan. Use `SHA256SUMS-v1.4.2` to verify every Python and runtime artifact after download.

**Full Changelog:** https://github.com/modelscope/FunASR/compare/v1.4.1...v1.4.2

<!-- funasr-runtime-downloads:start -->
## Runtime downloads

This Python release pairs with the current prebuilt llama.cpp / GGUF runtime release: [runtime-llamacpp-v0.2.0](https://github.com/modelscope/FunASR/releases/tag/runtime-llamacpp-v0.2.0).

The same verified runtime assets are attached directly to this Python release so users can find the package and self-contained `llama-funasr-*` binaries in one place.

| Platform | Asset | SHA-256 |
|---|---|---|
| Linux arm64 | [funasr-llamacpp-linux-arm64.tar.gz](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-linux-arm64.tar.gz) | `c78987b2384c6aef339aea1bcd0e130070455d6394fa7ab7ca26840ead10d5da` |
| Linux x64 AVX2 | [funasr-llamacpp-linux-x64-avx2.tar.gz](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-linux-x64-avx2.tar.gz) | `02e10e9a46ea76a040c45d431efe51a3324e64f08c24d38e18c8a4d2781490cd` |
| Linux x64 Vulkan | [funasr-llamacpp-linux-x64-vulkan.tar.gz](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-linux-x64-vulkan.tar.gz) | `caf71b8c0b4c3249cebc4175e5406d3c588eb9e8966a00d571d4cc5070405385` |
| Linux x64 portable | [funasr-llamacpp-linux-x64.tar.gz](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-linux-x64.tar.gz) | `15e6407143b4fb91d90bb37f2a41c64c4d48ea0fbe6404b88a9b70269c84f240` |
| macOS arm64 | [funasr-llamacpp-macos-arm64.tar.gz](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-macos-arm64.tar.gz) | `416cbb289e31cb7575365d382155074e922fd061807a37b9ca0247dabd9bc6f9` |
| Windows x64 AVX2 | [funasr-llamacpp-windows-x64-avx2.zip](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-windows-x64-avx2.zip) | `4db0f11f603c324a63545cd7009cdd45bb45576efe282cec22796b5fd42d8ea1` |
| Windows x64 CUDA | [funasr-llamacpp-windows-x64-cuda.zip](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-windows-x64-cuda.zip) | `7f2f9ef4d7e0291b284a295ec74bbeca9ea635a7f5f42d0ad06eb780c0d6efc1` |
| Windows x64 Vulkan | [funasr-llamacpp-windows-x64-vulkan.zip](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-windows-x64-vulkan.zip) | `90b45240c6ccc9177c25490a11848de60a406e129391c8736b14521c0c28cdcb` |
| Windows x64 portable | [funasr-llamacpp-windows-x64.zip](https://github.com/modelscope/FunASR/releases/download/v1.4.2/funasr-llamacpp-windows-x64.zip) | `297c962346d7e30d7a7c2c860dfaab3ff07d01fddf15e6fc5212ca9545441a51` |

Quick start: download one asset, unpack it, then run the bundled `download-funasr-model.sh <sensevoice|paraformer|nano>` helper and one of `llama-funasr-cli`, `llama-funasr-sensevoice`, or `llama-funasr-paraformer`.

For Python users, install from PyPI:

```bash
python -m pip install -U "funasr==1.4.2"
```
<!-- funasr-runtime-downloads:end -->