Merge f99e5fc706 into 5115a066c9

* Fix crash in ASR tasks when lm is set to none in #2237

* fix #2587: Resolve VAD multithreading issue

* Update funasr/models/fsmn_vad_streaming/model.py

ok

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

README.md

@@ -76,7 +76,7 @@
 ```text
 python>=3.8
 torch>=1.13
-torchaudio
+torchaudio==v2.4.0
 ```
 
 - Install for pypi

SECURITY.md

@@ -0,0 +1,11 @@
+# Security Policy
+
+## Reporting a Vulnerability
+
+If you believe you have discovered a security vulnerability, please report it to us through the following portal:
+
+👉[Report Security Issue](https://yundun.console.aliyun.com/?p=xznew#/taskmanagement/tasks/detail/153)
+
+> **Note:** This channel is strictly for reporting security-related issues. Non-security vulnerabilities or general bug reports will not be addressed here.
+
+We sincerely appreciate your responsible disclosure and your contribution to helping us keep our project secure.

funasr/auto/auto_model.py

@@ -301,14 +301,27 @@ class AutoModel:
         res = self.model(*args, kwargs)
         return res
 
-    def generate(self, input, input_len=None, **cfg):
+    def generate(self, input, input_len=None, progress_callback=None, **cfg):
         if self.vad_model is None:
-            return self.inference(input, input_len=input_len, **cfg)
+            return self.inference(
+                input, input_len=input_len, progress_callback=progress_callback, **cfg
+            )
 
         else:
-            return self.inference_with_vad(input, input_len=input_len, **cfg)
+            return self.inference_with_vad(
+                input, input_len=input_len, progress_callback=progress_callback, **cfg
+            )
 
-    def inference(self, input, input_len=None, model=None, kwargs=None, key=None, **cfg):
+    def inference(
+        self,
+        input,
+        input_len=None,
+        model=None,
+        kwargs=None,
+        key=None,
+        progress_callback=None,
+        **cfg,
+    ):
         kwargs = self.kwargs if kwargs is None else kwargs
         if "cache" in kwargs:
             kwargs.pop("cache")
@@ -365,6 +378,11 @@ class AutoModel:
             if pbar:
                 pbar.update(end_idx - beg_idx)
                 pbar.set_description(description)
+            if progress_callback:
+                try:
+                    progress_callback(end_idx, num_samples)
+                except Exception as e:
+                    logging.error(f"progress_callback error: {e}")
             time_speech_total += batch_data_time
             time_escape_total += time_escape
 

funasr/models/fsmn_vad_streaming/model.py

@@ -652,10 +652,11 @@ class FsmnVADStreaming(nn.Module):
         key: list = None,
         tokenizer=None,
         frontend=None,
-        cache: dict = {},
+        cache: dict = None,
         **kwargs,
     ):
-
+        if cache is None:
+            cache = {}
         if len(cache) == 0:
             self.init_cache(cache, **kwargs)
 

runtime/python/onnxruntime/funasr_onnx/utils/utils.py

@@ -221,10 +221,10 @@ class OrtInferSession:
                 RuntimeWarning,
             )
 
-    def __call__(self, input_content: List[Union[np.ndarray, np.ndarray]]) -> np.ndarray:
+    def __call__(self, input_content: List[Union[np.ndarray, np.ndarray]], run_options = None) -> np.ndarray:
         input_dict = dict(zip(self.get_input_names(), input_content))
         try:
-            return self.session.run(self.get_output_names(), input_dict)
+            return self.session.run(self.get_output_names(), input_dict, run_options)
         except Exception as e:
             raise ONNXRuntimeError("ONNXRuntime inferece failed.") from e
 

tests/test_auto_model.py

@@ -22,7 +22,39 @@ class TestAutoModel(unittest.TestCase):
         kwargs["spk_kwargs"] = {"cb_kwargs": {"merge_thr": merge_thr}}
         model = AutoModel(**kwargs)
         self.assertEqual(model.cb_model.model_config['merge_thr'], merge_thr)
-        # res = model.generate(input="/test.wav", 
+        # res = model.generate(input="/test.wav",
         #              batch_size_s=300)
+
+    def test_progress_callback_called(self):
+        class DummyModel:
+            def __init__(self):
+                self.param = torch.nn.Parameter(torch.zeros(1))
+
+            def parameters(self):
+                return iter([self.param])
+
+            def eval(self):
+                pass
+
+            def inference(self, data_in=None, **kwargs):
+                results = [{"text": str(d)} for d in data_in]
+                return results, {"batch_data_time": 1}
+
+        am = AutoModel.__new__(AutoModel)
+        am.model = DummyModel()
+        am.kwargs = {"batch_size": 2, "disable_pbar": True}
+
+        progress = []
+
+        res = AutoModel.inference(
+            am,
+            ["a", "b", "c"],
+            progress_callback=lambda idx, total: progress.append((idx, total)),
+        )
+
+        self.assertEqual(len(progress), 2)
+        self.assertEqual(progress, [(2, 3), (3, 3)])
+
+
 if __name__ == '__main__':
-    unittest.main()
+    unittest.main()