Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

update

游雁9fb473bc89f6
diff --git a/funasr/models/llm_asr/model.py b/funasr/models/llm_asr/model.py
index 937c8e1bf..2b6db96b8 100644
--- a/funasr/models/llm_asr/model.py
+++ b/funasr/models/llm_asr/model.py
@@ -216,8 +216,8 @@ class LLMASRNAR(nn.Module):
self, speech: torch.Tensor, speech_lengths: torch.Tensor, **kwargs,
) -> Tuple[torch.Tensor, torch.Tensor]:
- audio_mask = kwargs.get("audio_mask")
- audio_token_lengths = audio_mask.sum(-1) if audio_mask else None
+ audio_mask = kwargs.get("audio_mask", None)
+ audio_token_lengths = audio_mask.sum(-1) if audio_mask is not None else None
batch = {"speech": speech, "speech_lengths": speech_lengths}
enc, enc_lens = self.audio_encoder.encode(**batch)