From 4b0ad793b3159c0932376a0b45ada1cd155b2ba0 Mon Sep 17 00:00:00 2001
From: Yabin Li <wucong.lyb@alibaba-inc.com>
Date: 星期四, 07 三月 2024 15:19:56 +0800
Subject: [PATCH] Update CMakeLists.txt
---
funasr/models/llm_asr_nar/model.py | 7 +++++--
1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/funasr/models/llm_asr_nar/model.py b/funasr/models/llm_asr_nar/model.py
index 6a4ecce..0f4973d 100644
--- a/funasr/models/llm_asr_nar/model.py
+++ b/funasr/models/llm_asr_nar/model.py
@@ -214,7 +214,7 @@
def encode(
self, speech: torch.Tensor, speech_lengths: torch.Tensor, **kwargs,
- ) -> Tuple[torch.Tensor, torch.Tensor]:
+ ):
audio_mask = kwargs.get("audio_mask", None)
audio_token_lengths = audio_mask.sum(-1) if audio_mask is not None else None
@@ -315,7 +315,10 @@
model_outputs = self.llm(inputs_embeds=inputs_embeds, attention_mask=attention_mask, labels=None)
preds = torch.argmax(model_outputs.logits, -1)
text = tokenizer.batch_decode(preds, add_special_tokens=False, skip_special_tokens=True)
- text = text[0].split(': \n')[-1]
+
+ text = text[0].split(': ')[-1]
+ text = text.strip()
+
# preds = torch.argmax(model_outputs.logits, -1)
ibest_writer = None
--
Gitblit v1.9.1