Adds multi-language support for VITS onnx, fixes onnx inference error…

… when speaker_id is None or not passed, fixes onnx exporting for models with init_discriminator=false (coqui-ai#2816)

TTS/tts/models/vits.py

@@ -1814,14 +1814,16 @@ def export_onnx(self, output_path: str = "coqui_vits.onnx", verbose: bool = True
 
         # rollback values
         _forward = self.forward
-        disc = self.disc
+        disc = None
+        if hasattr(self, 'disc'):
+            disc = self.disc
         training = self.training
 
         # set export mode
         self.disc = None
         self.eval()
 
-        def onnx_inference(text, text_lengths, scales, sid=None):
+        def onnx_inference(text, text_lengths, scales, sid=None, langid=None):
             noise_scale = scales[0]
             length_scale = scales[1]
             noise_scale_dp = scales[2]
@@ -1834,7 +1836,7 @@ def onnx_inference(text, text_lengths, scales, sid=None):
                     "x_lengths": text_lengths,
                     "d_vectors": None,
                     "speaker_ids": sid,
-                    "language_ids": None,
+                    "language_ids": langid,
                     "durations": None,
                 },
             )["model_outputs"]
@@ -1845,11 +1847,14 @@ def onnx_inference(text, text_lengths, scales, sid=None):
         dummy_input_length = 100
         sequences = torch.randint(low=0, high=self.args.num_chars, size=(1, dummy_input_length), dtype=torch.long)
         sequence_lengths = torch.LongTensor([sequences.size(1)])
-        sepaker_id = None
+        speaker_id = None
+        language_id = None
         if self.num_speakers > 1:
-            sepaker_id = torch.LongTensor([0])
+            speaker_id = torch.LongTensor([0])
+        if self.num_languages > 0 and self.embedded_language_dim > 0:
+            language_id = torch.LongTensor([0])
         scales = torch.FloatTensor([self.inference_noise_scale, self.length_scale, self.inference_noise_scale_dp])
-        dummy_input = (sequences, sequence_lengths, scales, sepaker_id)
+        dummy_input = (sequences, sequence_lengths, scales, speaker_id, language_id)
 
         # export to ONNX
         torch.onnx.export(
@@ -1858,7 +1863,7 @@ def onnx_inference(text, text_lengths, scales, sid=None):
             opset_version=15,
             f=output_path,
             verbose=verbose,
-            input_names=["input", "input_lengths", "scales", "sid"],
+            input_names=["input", "input_lengths", "scales", "sid", "langid"],
             output_names=["output"],
             dynamic_axes={
                 "input": {0: "batch_size", 1: "phonemes"},
@@ -1871,7 +1876,8 @@ def onnx_inference(text, text_lengths, scales, sid=None):
         self.forward = _forward
         if training:
             self.train()
-        self.disc = disc
+        if not disc is None:
+            self.disc = disc
 
     def load_onnx(self, model_path: str, cuda=False):
         import onnxruntime as ort
@@ -1888,7 +1894,7 @@ def load_onnx(self, model_path: str, cuda=False):
             providers=providers,
         )
 
-    def inference_onnx(self, x, x_lengths=None, speaker_id=None):
+    def inference_onnx(self, x, x_lengths=None, speaker_id=None, language_id=None):
         """ONNX inference"""
 
         if isinstance(x, torch.Tensor):
@@ -1903,13 +1909,15 @@ def inference_onnx(self, x, x_lengths=None, speaker_id=None):
             [self.inference_noise_scale, self.length_scale, self.inference_noise_scale_dp],
             dtype=np.float32,
         )
+
         audio = self.onnx_sess.run(
             ["output"],
             {
                 "input": x,
                 "input_lengths": x_lengths,
                 "scales": scales,
-                "sid": torch.tensor([speaker_id]).cpu().numpy(),
+                "sid": None if speaker_id is None else torch.tensor([speaker_id]).cpu().numpy(),
+				"langid": None if language_id is None else torch.tensor([language_id]).cpu().numpy()
             },
         )
         return audio[0][0]