Update to v2

README.md

@@ -23,16 +23,23 @@
         │   └── README.md
         ├── models
         │   ├── am.mvn
-        │   ├── asr_paraformer.onnx  # 放在这里
+        │   ├── asr_paraformerv2.onnx  # 放在这里
         │   └── token_list.pkl
         ├── rapid_paraformer.py
         └── utils.py
         ```
 
 3. 运行demo
-   ```bash
-   python demo.py
-   ```
+    ```python
+    from rapid_paraformer import RapidParaformer
+
+    paraformer = RapidParaformer()
+
+    wav_path = 'test_wavs/example_test.wav'
+    print(wav_path)
+    result = paraformer(str(wav_path))
+    print(result)
+    ```
 4. 查看结果
    ```text
    [['呃说不配合就不配合的好以上的话呢我们摘取八九十三条因为这三条的话呢比较典型啊一些数字比较明确尤其是时间那么我们要投资者就是了解这一点啊不要轻信这个市场可以快速回来啊这些配市公司啊后期又利好了可

demo.py

@@ -6,6 +6,7 @@ from rapid_paraformer import RapidParaformer
 
 paraformer = RapidParaformer()
 
-wav_file = 'test_wavs/0478_00017.wav'
-result = paraformer(wav_file)
+wav_path = 'test_wavs/example_test.wav'
+print(wav_path)
+result = paraformer(str(wav_path))
 print(result)

rapid_paraformer/config.yaml

@@ -20,7 +20,7 @@ WavFrontend:
     filter_length_max: -.inf
 
 Model:
-  model_path: models/asr_paraformer.onnx
+  model_path: models/asr_paraformerv2.onnx
   use_cuda: false
   CUDAExecutionProvider:
       device_id: 0

rapid_paraformer/utils.py

@@ -195,7 +195,7 @@ class WavFrontend():
             feats_lens.append(mat.shape[0])
 
         feats_pad = np.array(feats).astype(np.float32)
-        feats_lens = np.array(feats_lens).astype(np.int64)
+        feats_lens = np.array(feats_lens).astype(np.int32)
         return feats_pad, feats_lens
 
     @staticmethod
@@ -214,7 +214,7 @@ class WavFrontend():
             else:
                 # process last LFR frame
                 num_padding = lfr_m - (T - i * lfr_n)
-                frame = (inputs[i * lfr_n:]).view(-1)
+                frame = inputs[i * lfr_n:].reshape(-1)
                 for _ in range(num_padding):
                     frame = np.hstack((frame, inputs[-1]))
 
@@ -320,8 +320,8 @@ class OrtInferSession():
     def get_input_names(self, ):
         return [v.name for v in self.session.get_inputs()]
 
-    def get_output_name(self, output_idx=0):
-        return self.session.get_outputs()[output_idx].name
+    def get_output_names(self,):
+        return [v.name for v in self.session.get_outputs()]
 
     def get_character_list(self, key: str = 'character'):
         return self.meta_dict[key].splitlines()