From b3bc53e33264e5b289d2e044859dd001213b207e Mon Sep 17 00:00:00 2001
From: shixian.shi <shixian.shi@alibaba-inc.com>
Date: 星期四, 11 五月 2023 15:50:17 +0800
Subject: [PATCH] update timestamp for batch inference

---
 funasr/bin/asr_inference_paraformer_vad_punc.py |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)

diff --git a/funasr/bin/asr_inference_paraformer_vad_punc.py b/funasr/bin/asr_inference_paraformer_vad_punc.py
index 8ecba32..8555449 100644
--- a/funasr/bin/asr_inference_paraformer_vad_punc.py
+++ b/funasr/bin/asr_inference_paraformer_vad_punc.py
@@ -642,7 +642,16 @@
             # result = result_segments[0]
             text, token, token_int = result[0], result[1], result[2]
             time_stamp = None if len(result) < 5 else result[4]
-
+            # post process timestamp
+            offset_list = [0]
+            offset = 0
+            for i in range(1, len(time_stamp)):
+                if time_stamp[i][0] < time_stamp[i-1][1]:
+                    offset += time_stamp[i-1][1]
+                offset_list.append(offset)
+            for i in range(1, len(time_stamp)):
+                time_stamp[i][0] += offset_list[i]
+                time_stamp[i][1] += offset_list[i]
 
             if use_timestamp and time_stamp is not None:
                 postprocessed_result = postprocess_utils.sentence_postprocess(token, time_stamp)

--
Gitblit v1.9.1