From 59bc02b089f7a626fe67907dcfc695eae6883f82 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期五, 14 六月 2024 13:59:49 +0800
Subject: [PATCH] decoding
---
funasr/datasets/openai_datasets/datasets.py | 11 ++++++++---
1 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/funasr/datasets/openai_datasets/datasets.py b/funasr/datasets/openai_datasets/datasets.py
index 39b8453..7300b9d 100644
--- a/funasr/datasets/openai_datasets/datasets.py
+++ b/funasr/datasets/openai_datasets/datasets.py
@@ -64,6 +64,8 @@
self.max_token_length = kwargs.get("max_token_length", 1024)
self.batch_size_scale_ratio_max = kwargs.get("batch_size_scale_ratio_max", 1.5)
self.batch_size_token_max = kwargs.get("batch_size_token_max", 2500)
+ self.audio_adaptor_downsample_rate = kwargs.get("audio_adaptor_downsample_rate", 2)
+ self.audio_encoder_downsample_rate = kwargs.get("audio_encoder_downsample_rate", 4)
def get_source_len(self, index):
item = self.index_ds[index]
@@ -136,10 +138,13 @@
speech = speech.permute(0, 2, 1)
# if speech_lengths > self.batch_size:
# continue
+ if self.audio_encoder_downsample_rate == 4:
+ olens = 1 + (speech_lengths[0].item() - 3 + 2 * 1) // 2
+ olens = 1 + (olens - 3 + 2 * 1) // 2
+ elif self.audio_encoder_downsample_rate == 1:
+ olens = speech_lengths[0].item()
- olens = 1 + (speech_lengths[0].item() - 3 + 2 * 1) // 2
- olens = 1 + (olens - 3 + 2 * 1) // 2
- sub_token_len = (olens - 1) // 2 + 1
+ sub_token_len = (olens - 1) // self.audio_adaptor_downsample_rate + 1
sub_token = [0] * sub_token_len
fbank_beg_i = [len(source_ids)]
source_ids += sub_token
--
Gitblit v1.9.1