From c5acc04e2df3316c284c3ab75575498934314560 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期四, 30 三月 2023 16:35:03 +0800
Subject: [PATCH] Merge branch 'dev_cmz2' of github.com:alibaba-damo-academy/FunASR into dev_cmz2 add

---
 funasr/export/models/vad_realtime_transformer.py |   23 +++++++++++++----------
 1 files changed, 13 insertions(+), 10 deletions(-)

diff --git a/funasr/export/models/vad_realtime_transformer.py b/funasr/export/models/vad_realtime_transformer.py
index 44583d8..de7c721 100644
--- a/funasr/export/models/vad_realtime_transformer.py
+++ b/funasr/export/models/vad_realtime_transformer.py
@@ -21,7 +21,9 @@
         **kwargs,
     ):
         super().__init__()
-
+        onnx = False
+        if "onnx" in kwargs:
+            onnx = kwargs["onnx"]
 
         self.embed = model.embed
         if isinstance(model.encoder, SANMVadEncoder):
@@ -30,6 +32,7 @@
             assert False, "Only support samn encode."
         # self.encoder = model.encoder
         self.decoder = model.decoder
+        self.model_name = model_name
 
 
 
@@ -44,7 +47,7 @@
         """
         x = self.embed(input)
         # mask = self._target_mask(input)
-        h, _, _ = self.encoder(x, text_lengths, vad_indexes)
+        h, _ = self.encoder(x, text_lengths, vad_indexes)
         y = self.decoder(h)
         return y
 
@@ -53,12 +56,13 @@
 
     def get_dummy_inputs(self):
         length = 120
-        text_indexes = torch.randint(0, self.embed.num_embeddings, (2, length))
-        text_lengths = torch.tensor([length-20, length], dtype=torch.int32)
-        return (text_indexes, text_lengths)
+        text_indexes = torch.randint(0, self.embed.num_embeddings, (1, length))
+        text_lengths = torch.tensor([length], dtype=torch.int32)
+        vad_mask = torch.ones(length, length, dtype=torch.float32)[None, None, :, :]
+        return (text_indexes, text_lengths, vad_mask)
 
     def get_input_names(self):
-        return ['input', 'text_lengths']
+        return ['input', 'text_lengths', 'vad_mask']
 
     def get_output_names(self):
         return ['logits']
@@ -66,14 +70,13 @@
     def get_dynamic_axes(self):
         return {
             'input': {
-                0: 'batch_size',
                 1: 'feats_length'
             },
-            'text_lengths': {
-                0: 'batch_size',
+            'vad_mask': {
+                2: 'feats_length1',
+                3: 'feats_length2'
             },
             'logits': {
-                0: 'batch_size',
                 1: 'logits_length'
             },
         }

--
Gitblit v1.9.1