From fe588bc508c0076bb007d6ed36c18ac8ecb341ac Mon Sep 17 00:00:00 2001
From: 王梦迪 <73778524+di-osc@users.noreply.github.com>
Date: 星期二, 20 五月 2025 16:10:59 +0800
Subject: [PATCH] Fsmn_vad支持多线程并发调用 (#2519)
---
runtime/python/onnxruntime/funasr_onnx/vad_bin.py | 24 ++++++++++++++++++++----
1 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/runtime/python/onnxruntime/funasr_onnx/vad_bin.py b/runtime/python/onnxruntime/funasr_onnx/vad_bin.py
index 92928a8..af4663a 100644
--- a/runtime/python/onnxruntime/funasr_onnx/vad_bin.py
+++ b/runtime/python/onnxruntime/funasr_onnx/vad_bin.py
@@ -69,7 +69,7 @@
model_file, device_id, intra_op_num_threads=intra_op_num_threads
)
self.batch_size = batch_size
- self.vad_scorer = E2EVadModel(config["model_conf"])
+ self.vad_scorer_config = config["model_conf"]
self.max_end_sil = (
max_end_sil if max_end_sil is not None else config["model_conf"]["max_end_silence_time"]
)
@@ -90,10 +90,9 @@
waveform_list = self.load_data(audio_in, self.frontend.opts.frame_opts.samp_freq)
waveform_nums = len(waveform_list)
is_final = kwargs.get("kwargs", False)
-
segments = [[]] * self.batch_size
for beg_idx in range(0, waveform_nums, self.batch_size):
-
+ vad_scorer = E2EVadModel(self.vad_scorer_config)
end_idx = min(waveform_nums, beg_idx + self.batch_size)
waveform = waveform_list[beg_idx:end_idx]
feats, feats_len = self.extract_feat(waveform)
@@ -122,7 +121,7 @@
inputs.extend(in_cache)
scores, out_caches = self.infer(inputs)
in_cache = out_caches
- segments_part = self.vad_scorer(
+ segments_part = vad_scorer(
scores,
waveform_package,
is_final=is_final,
@@ -143,7 +142,24 @@
return segments
def load_data(self, wav_content: Union[str, np.ndarray, List[str]], fs: int = None) -> List:
+
+ def convert_to_wav(input_path, output_path):
+ from pydub import AudioSegment
+ try:
+ audio = AudioSegment.from_mp3(input_path)
+ audio.export(output_path, format="wav")
+ print("闊抽鏂囦欢涓簃p3鏍煎紡锛屽凡杞崲涓簑av鏍煎紡")
+
+ except Exception as e:
+ print(f"杞崲澶辫触:{e}")
+
def load_wav(path: str) -> np.ndarray:
+ if not path.lower().endswith('.wav'):
+ import os
+ input_path = path
+ path = os.path.splitext(path)[0]+'.wav'
+ convert_to_wav(input_path,path) #灏唌p3鏍煎紡杞崲鎴恮av鏍煎紡
+
waveform, _ = librosa.load(path, sr=fs)
return waveform
--
Gitblit v1.9.1