From e8f80e96f99cb856423d030c7d055c302a6d3278 Mon Sep 17 00:00:00 2001
From: zhifu gao <zhifu.gzf@alibaba-inc.com>
Date: 星期三, 17 四月 2024 21:23:20 +0800
Subject: [PATCH] Dev gzf exp (#1627)
---
funasr/datasets/audio_datasets/index_ds.py | 31 ++++++++++++++++++++++---------
1 files changed, 22 insertions(+), 9 deletions(-)
diff --git a/funasr/datasets/audio_datasets/index_ds.py b/funasr/datasets/audio_datasets/index_ds.py
index 34f7b4f..53419e8 100644
--- a/funasr/datasets/audio_datasets/index_ds.py
+++ b/funasr/datasets/audio_datasets/index_ds.py
@@ -76,7 +76,10 @@
def __init__(self, path: str, **kwargs):
super().__init__()
-
+ self.max_source_length = kwargs.get("max_source_length", 2048)
+ self.min_source_length = kwargs.get("min_source_length", 0)
+ self.max_target_length = kwargs.get("max_target_length", 2048)
+ self.min_target_length = kwargs.get("min_target_length", 0)
if isinstance(path, (list, tuple)): # wav.scp, text.txt/text.trans
from funasr.datasets.audio_datasets.scp2jsonl import gen_jsonl_from_wav_text_list
jsonl_outdir = os.path.dirname(path[0])
@@ -92,7 +95,7 @@
for line in fin:
data = json.loads(line.strip())
if "text" in data: # for sft
- self.contents.append(data['text'])
+ contents.append(data['text'])
if "source" in data: # for speech lab pretrain
prompt = data.get("prompt", "<ASR>")
source = data["source"]
@@ -101,13 +104,23 @@
target_len = data.get("target_len", 0)
if "aishell" in source:
target = target.replace(" ", "")
- contents.append({"source": source,
- "prompt": prompt,
- "target": target,
- "source_len": source_len,
- "target_len": target_len,
- }
- )
+ if source_len < self.min_source_length or source_len > self.max_source_length:
+ continue
+ if target_len < self.min_target_length or target_len > self.max_target_length:
+ continue
+ contents_i = {"source": source,
+ "prompt": prompt,
+ "target": target,
+ "source_len": source_len,
+ "target_len": target_len,
+ }
+ text_language = data.get("text_language", None)
+ if text_language is not None:
+ contents_i["text_language"] = text_language
+ audio_language = data.get("audio_language", None)
+ if audio_language is not None:
+ contents_i["audio_language"] = audio_language
+ contents.append(contents_i)
self.contents = contents
--
Gitblit v1.9.1