bugfix

游雁

2024-04-15 199f6e630af9d9ccfb8c50bcd1f7ba0392c9bd5e

 funasr/models/campplus/model.py

@@ -123,5 +123,5 @@
        time3 = time.perf_counter()
        meta_data["extract_feat"] = f"{time3 - time2:0.3f}"
        meta_data["batch_data_time"] = np.array(speech_times).sum().item() / 16000.0
        results = [{"spk_embedding": self.forward(speech)}]
        results = [{"spk_embedding": self.forward(speech.to(torch.float32))}]
        return results, meta_data