Instructions to use AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0 # Run inference directly in the terminal: llama cli -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0 # Run inference directly in the terminal: llama cli -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0 # Run inference directly in the terminal: ./build/bin/llama cli -hf AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
Use Docker
docker model run hf.co/AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
- Llama new
- LM Studio
- Jan
- Ollama
How to use AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp with Ollama:
ollama run hf.co/AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
- Unsloth Desktop
- Docker Model Runner
How to use AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp with Docker Model Runner:
docker model run hf.co/AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
- Lemonade
How to use AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp:Q8_0
Run and chat with the model
lemonade run user.Qwen3-ASR-1.7B-AX8860-llama.cpp-Q8_0
List all available models
lemonade list
- Atomic Chat
Qwen3-ASR-1.7B on AX8860 (llama.cpp + ggml-axera, 2026-11-01)
在爱芯 AX8860 的 NPU(4 核)上用 llama.cpp 运行 Qwen3-ASR-1.7B 语音识别。 Qwen3-ASR-1.7B speech recognition on the 4-core NPU of the Axera AX8860, with llama.cpp.
文件 / Files
| 文件 / File | 说明 / Description |
|---|---|
qwen3-asr-1.7b-bf16.gguf |
语言模型权重(GGUF BF16,3.2 GB,md5 d1a6720b651a008ccb99dbde8b1b06ce)/ language model weights |
mmproj-qwen3-asr-1.7b-q8_0.gguf |
音频编码器权重(GGUF Q8_0,339 MB,md5 8c288771eb3b79405229f9aef93729cd)/ audio encoder weights |
bin/ |
llama.cpp 板端程序 llama-mtmd-cli、llama-server(静态链接,只依赖系统 /opt/lib 与 libc)/ the board programs (static; need only the system /opt/lib and libc) |
kernels/ |
NPU kernel:音频编码器、文本 prefill、文本 decode(含输出头)各一个 + manifest.json / one kernel each for the audio encoder, the text prefill and the text decode (with the output head) + manifest.json |
run.sh |
运行脚本 / run script |
权重只来自 GGUF:kernel 不含权重,加载时从 GGUF 转换(语言模型 MXFP8,音频编码器 BF16:mmproj 中为 Q8_0,加载时反量化)。 The weights come only from the GGUFs: the kernels hold none; they are converted at load (language model MXFP8, audio encoder BF16, dequantized from the mmproj's Q8_0).
运行 / Run
环境 / Requirements:AX8860,SDK V0.5.0 或更新(cat /proc/ax_proc/version)/ SDK V0.5.0 or newer.
./run.sh speech.wav # 转写一段录音 / transcribe one recording
./run.sh -s 8080 # OpenAI 兼容服务 / OpenAI-compatible server
服务请求 / Server request:/v1/chat/completions,content 里放 / put in the content
{"type":"input_audio","input_audio":{"format":"wav","data":"<base64>"}。
直接调用程序时设置 / When calling the programs directly: LD_LIBRARY_PATH=/opt/lib GGML_AXERA_KERNEL_DIR=<repo>/kernels,
-ngl 99,-c 最大 4096;服务端 -np 1。
性能 / Performance
AX8860,SDK V0.5.1,NPU 4 核,服务模式端到端(请求发出到收到结果)/ 4 cores, server, end to end:
| 录音 / Recording | 耗时 / Time | RTF |
|---|---|---|
| 0.6 s | 0.10 s | 0.17 |
| 7.3 s | 0.53–0.64 s | 0.07–0.09 |
| 60 s | 2.5 s | 0.041 |
首次请求另加约 0.5 s 的权重转换(每次启动一次)/ the first request adds ~0.5 s of weight conversion (once per start).
本版 decode kernel 直接读 prefill 的权重 bank,NPU 内存里只放一份层权重:加载后 NPU 内存(CMM)6035 → 5111 MB,输出不变,decode 略快。 This release's decode kernel reads the prefill weight banks (one copy of the layer weights in NPU memory): CMM after load 6035 -> 5111 MB, same outputs, slightly faster decoding.
正确性 / Correctness
测试录音的转写与上一版 NPU 实现(20 句评测与参考逐字一致)逐字相同。 The test recordings' transcripts are identical to those of the previous NPU implementation (20/20 sentences exact on its evaluation set).
已知限制 / Known limits
- 运行时独占 NPU 4 核 / takes all 4 NPU cores; do not run other NPU programs alongside.
- 单序列:服务端须
-np 1;上下文最大 4096 / one sequence: the server needs-np 1; context up to 4096.
源码 / Sources
- llama.cpp:ZHEQIUSHUI/llama.cpp
axera-dev@57a30bd(ggml/src/ggml-axera/models/qwen3、qwen3_asr) - kernel:ZHEQIUSHUI/ax-dsl-kernels
main@804a260(models/qwen3_asr)
- Downloads last month
- 279
16-bit
Model tree for AXERA-TECH/Qwen3-ASR-1.7B-AX8860-llama.cpp
Base model
Qwen/Qwen3-ASR-1.7B