From d77910eb6d171727f2350e45c31c91436c4c8891 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期一, 11 十二月 2023 13:42:40 +0800
Subject: [PATCH] funasr2
---
funasr/datasets/dataset_jsonl.py | 20 ++++++++++++++------
1 files changed, 14 insertions(+), 6 deletions(-)
diff --git a/funasr/datasets/dataset_jsonl.py b/funasr/datasets/dataset_jsonl.py
index eef67c5..7f2cd83 100644
--- a/funasr/datasets/dataset_jsonl.py
+++ b/funasr/datasets/dataset_jsonl.py
@@ -6,6 +6,7 @@
import librosa
import torchaudio
import time
+import logging
def load_audio(audio_path: str, fs: int=16000):
audio = None
@@ -41,8 +42,7 @@
def __init__(self, path):
super().__init__()
- # data_parallel_size = dist.get_world_size()
- data_parallel_size = 1
+
contents = []
with open(path, encoding='utf-8') as fin:
for line in fin:
@@ -66,12 +66,20 @@
self.contents = []
total_num = len(contents)
- num_per_rank = total_num // data_parallel_size
- # rank = dist.get_rank()
- rank = 0
+ try:
+ rank = dist.get_rank()
+ world_size = dist.get_world_size()
+ except:
+ rank = 0
+ world_size = 1
+ logging.warning("distributed is not initialized, only single shard")
+ num_per_rank = total_num // world_size
+
+ # rank = 0
# import ipdb; ipdb.set_trace()
self.contents = contents[rank * num_per_rank:(rank + 1) * num_per_rank]
-
+
+ logging.info("in rank: {}, num of samplers: {}, total_num of samplers across ranks: {}".format(rank, len(self.contents), len(contents)))
def __len__(self):
return len(self.contents)
--
Gitblit v1.9.1