From 91027ddab49e5791fc42569b4db9dafca55735e6 Mon Sep 17 00:00:00 2001
From: 凌匀 <ailsa.zly@alibaba-inc.com>
Date: 星期四, 16 二月 2023 22:11:18 +0800
Subject: [PATCH] fix vad results bug
---
funasr/export/models/e2e_asr_paraformer.py | 9 +++------
1 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/funasr/export/models/e2e_asr_paraformer.py b/funasr/export/models/e2e_asr_paraformer.py
index 8388f4f..bf5ed1e 100644
--- a/funasr/export/models/e2e_asr_paraformer.py
+++ b/funasr/export/models/e2e_asr_paraformer.py
@@ -59,16 +59,13 @@
enc, enc_len = self.encoder(**batch)
mask = self.make_pad_mask(enc_len)[:, None, :]
pre_acoustic_embeds, pre_token_length, alphas, pre_peak_index = self.predictor(enc, mask)
- pre_token_length = pre_token_length.round().long()
+ pre_token_length = pre_token_length.round().type(torch.int32)
decoder_out, _ = self.decoder(enc, enc_len, pre_acoustic_embeds, pre_token_length)
decoder_out = torch.log_softmax(decoder_out, dim=-1)
- sample_ids = decoder_out.argmax(dim=-1)
+ # sample_ids = decoder_out.argmax(dim=-1)
- return decoder_out, sample_ids
-
- # def get_output_size(self):
- # return self.model.encoders[0].size
+ return decoder_out, pre_token_length
def get_dummy_inputs(self):
speech = torch.randn(2, 30, self.feats_dim)
--
Gitblit v1.9.1