Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
docs: add vLLM API repeat troubleshooting
zhifu gao0573e013d338
diff --git a/docs/vllm_guide.md b/docs/vllm_guide.mdindex 056798695..1b3d9b576 100644--- a/docs/vllm_guide.md+++ b/docs/vllm_guide.md@@ -309,6 +309,31 @@ for result in engine.streaming_generate("audio.wav", language="中文"):> **Note: `repetition_penalty` cannot be used with EmbedsPrompt.** Here the prompt is a sequence of embedding vectors with no corresponding token IDs, whereas `repetition_penalty` needs the prompt's token IDs to down-weight already-seen tokens in the logits; applying it under EmbedsPrompt **indexes out of bounds and triggers a CUDA device-side assert**.+### Production API Stability Checklist++When wrapping `AutoModelVLLM` in a long-running API service, keep request state isolated and pin safe decoding defaults:++```python+common = dict(+ language="auto",+ temperature=0.0,+ repetition_penalty=1.0,+ max_new_tokens=200,+)++for _ in range(2):+ results = model.generate(["vad_segment_01.wav", "vad_segment_02.wav"], **common)+ print([r["text"] for r in results])+```++If the same audio is normal on the first request but repeats on the second request:++1. Run the minimal script above outside the API layer with the same VAD segments.+2. If the script is stable, check whether the API wrapper reuses per-request variables, previous VAD segment lists, previous `results`, or accumulated text across requests.+3. If the script also repeats, capture the exact `funasr`, `vllm`, and `torch` versions, plus the first and second outputs, before tuning any decoding parameter.++Do not increase `repetition_penalty` to suppress repeats on Fun-ASR-Nano vLLM. The prompt-embeds path should stay at the neutral value `1.0`.+---## 5. Offline Speech Recognition Servicediff --git a/docs/vllm_guide_zh.md b/docs/vllm_guide_zh.mdindex a502df415..d453697ca 100644--- a/docs/vllm_guide_zh.md+++ b/docs/vllm_guide_zh.md@@ -304,6 +304,31 @@ for result in engine.streaming_generate("audio.wav", language="中文"):**注意:EmbedsPrompt 下不能用 `repetition_penalty`。** 此时 prompt 是 embedding 向量、没有对应的 token ID,而 `repetition_penalty` 要靠 prompt 的 token ID 在 logits 上给已出现的词降分;用在 EmbedsPrompt 上会**索引越界、触发 CUDA device-side assert**。+### 生产 API 稳定性清单++把 `AutoModelVLLM` 封装成长驻 API 服务时,请隔离每次请求的状态,并固定安全的解码默认值:++```python+common = dict(+ language="auto",+ temperature=0.0,+ repetition_penalty=1.0,+ max_new_tokens=200,+)++for _ in range(2):+ results = model.generate(["vad_segment_01.wav", "vad_segment_02.wav"], **common)+ print([r["text"] for r in results])+```++如果同一个音频第一次请求正常、第二次请求开始重复:++1. 先把 API 层拿掉,用相同 VAD 分段跑上面的最小脚本。+2. 如果最小脚本稳定,优先检查 API 封装是否复用了请求级变量、上一轮 VAD 分段列表、上一轮 `results` 或累积文本。+3. 如果最小脚本也重复,再记录完整的 `funasr`、`vllm`、`torch` 版本,以及第一次和第二次输出文本,再调整其它解码参数。++不要通过调大 `repetition_penalty` 来压制 Fun-ASR-Nano vLLM 重复输出;prompt-embeds 路径应保持中性值 `1.0`。+### 输出特性| 累积音频 | 输出质量 |diff --git a/docs/vllm_guide_zh_v2.md b/docs/vllm_guide_zh_v2.mdindex a502df415..d453697ca 100644--- a/docs/vllm_guide_zh_v2.md+++ b/docs/vllm_guide_zh_v2.md@@ -304,6 +304,31 @@ for result in engine.streaming_generate("audio.wav", language="中文"):**注意:EmbedsPrompt 下不能用 `repetition_penalty`。** 此时 prompt 是 embedding 向量、没有对应的 token ID,而 `repetition_penalty` 要靠 prompt 的 token ID 在 logits 上给已出现的词降分;用在 EmbedsPrompt 上会**索引越界、触发 CUDA device-side assert**。+### 生产 API 稳定性清单++把 `AutoModelVLLM` 封装成长驻 API 服务时,请隔离每次请求的状态,并固定安全的解码默认值:++```python+common = dict(+ language="auto",+ temperature=0.0,+ repetition_penalty=1.0,+ max_new_tokens=200,+)++for _ in range(2):+ results = model.generate(["vad_segment_01.wav", "vad_segment_02.wav"], **common)+ print([r["text"] for r in results])+```++如果同一个音频第一次请求正常、第二次请求开始重复:++1. 先把 API 层拿掉,用相同 VAD 分段跑上面的最小脚本。+2. 如果最小脚本稳定,优先检查 API 封装是否复用了请求级变量、上一轮 VAD 分段列表、上一轮 `results` 或累积文本。+3. 如果最小脚本也重复,再记录完整的 `funasr`、`vllm`、`torch` 版本,以及第一次和第二次输出文本,再调整其它解码参数。++不要通过调大 `repetition_penalty` 来压制 Fun-ASR-Nano vLLM 重复输出;prompt-embeds 路径应保持中性值 `1.0`。+### 输出特性| 累积音频 | 输出质量 |