Merge branch 'alibaba-damo-academy:main' into main

funasr/bin/vad_inference.py

@@ -109,7 +109,7 @@ class Speech2VadSegment:
             fbanks, fbanks_len = self.frontend.forward_fbank(speech, speech_lengths)
             feats, feats_len = self.frontend.forward_lfr_cmvn(fbanks, fbanks_len)
             fbanks = to_device(fbanks, device=self.device)
-            # feats = to_device(feats, device=self.device)
+            feats = to_device(feats, device=self.device)
             feats_len = feats_len.int()
         else:
             raise Exception("Need to extract feats first, please configure frontend configuration")
@@ -131,7 +131,7 @@ class Speech2VadSegment:
                 "in_cache": in_cache
             }
             # a. To device
-            batch = to_device(batch, device=self.device)
+            #batch = to_device(batch, device=self.device)
             segments_part, in_cache = self.vad_model(**batch)
             if segments_part:
                 for batch_num in range(0, self.batch_size):

funasr/models/frontend/wav_frontend.py

@@ -34,7 +34,7 @@ def load_cmvn(cmvn_file):
     means = np.array(means_list).astype(np.float)
     vars = np.array(vars_list).astype(np.float)
     cmvn = np.array([means, vars])
-    cmvn = torch.as_tensor(cmvn, dype=torch.float32)
+    cmvn = torch.as_tensor(cmvn, dtype=torch.float32)
     return cmvn
 
 

funasr/runtime/python/benchmark_onnx.md

@@ -1,4 +1,4 @@
-# CPU Benchmark (ONNX)
+# CPU Benchmark (ONNX-python)
 
 ## Configuration
 ### Data set:

funasr/version.txt

@@ -1 +1 @@
-0.4.2
+0.4.3