From 6e69d784e4814c3dbe35e8f70c6cf4b920c8b20b Mon Sep 17 00:00:00 2001
From: 天地 <tiandiweizun@gmail.com>
Date: 星期三, 19 三月 2025 23:10:13 +0800
Subject: [PATCH] 1. bug fix:list(mean)和list(var)，由于mean和var是numpy，导致写入到文件的格式错误，参考上面的话，大概率是list(mean.tolist())，其实外层list没有必要 (#2437)
---
funasr/datasets/audio_datasets/samplers.py | 9 +++------
1 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/funasr/datasets/audio_datasets/samplers.py b/funasr/datasets/audio_datasets/samplers.py
index bddf186..f7057de 100644
--- a/funasr/datasets/audio_datasets/samplers.py
+++ b/funasr/datasets/audio_datasets/samplers.py
@@ -392,17 +392,14 @@
)
batch = []
max_len_in_batch = 0
- count = 0
+ count = 1
for idx in buffer:
original_sample_length = self.dataset.get_source_len(idx)
if original_sample_length > self.max_token_length:
continue
sample_length = 1 if self.batch_type == "example" else original_sample_length
potential_batch_length = max(max_len_in_batch, sample_length) * (len(batch) + 1)
- if (
- potential_batch_length <= self.batch_size
- and count <= self.batch_size_sample_max
- ):
+ if potential_batch_length <= self.batch_size and count < self.batch_size_sample_max:
batch.append(idx)
max_len_in_batch = max(max_len_in_batch, sample_length)
count += 1
@@ -410,7 +407,7 @@
buffer_batches.append(batch)
batch = [idx]
max_len_in_batch = sample_length
- count = 0
+ count = 1
if batch:
buffer_batches.append(batch)
--
Gitblit v1.9.1