From 1596f6f414f6f41da66506debb1dff19fffeb3ec Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期一, 24 六月 2024 11:55:17 +0800
Subject: [PATCH] fixbug hotwords
---
funasr/datasets/large_datasets/build_dataloader.py | 32 ++++++++++++++++++++++----------
1 files changed, 22 insertions(+), 10 deletions(-)
diff --git a/funasr/datasets/large_datasets/build_dataloader.py b/funasr/datasets/large_datasets/build_dataloader.py
index 8a255f9..da04717 100644
--- a/funasr/datasets/large_datasets/build_dataloader.py
+++ b/funasr/datasets/large_datasets/build_dataloader.py
@@ -13,6 +13,7 @@
from funasr.register import tables
+
def read_symbol_table(symbol_table_file):
if isinstance(symbol_table_file, str):
symbol_table = {}
@@ -62,6 +63,7 @@
self._build_sentence_piece_processor()
return self.sp.DecodePieces(list(tokens))
+
@tables.register("dataset_classes", "LargeDataset")
class LargeDataLoader(AbsIterFactory):
def __init__(self, args, mode="train"):
@@ -76,22 +78,32 @@
bpe_tokenizer = SentencepiecesTokenizer(args.bpemodel)
self.dataset_conf = args.dataset_conf
if "frontend_conf" not in args:
- self.frontend_conf = None
+ self.frontend_conf = None
else:
self.frontend_conf = args.frontend_conf
- self.speed_perturb = args.speed_perturb if hasattr(args, "speed_perturb") else None
+ self.speed_perturb = args.speed_perturb if hasattr(args, "speed_perturb") else None
logging.info("dataloader config: {}".format(self.dataset_conf))
batch_mode = self.dataset_conf.get("batch_mode", "padding")
data_list = args.train_data_file if mode == "train" else args.valid_data_file
- self.dataset = Dataset(data_list, symbol_table, seg_dict, punc_dict, bpe_tokenizer,
- self.dataset_conf, self.frontend_conf,
- speed_perturb=self.speed_perturb if mode == "train" else None,
- mode=mode, batch_mode=batch_mode)
+ self.dataset = Dataset(
+ data_list,
+ symbol_table,
+ seg_dict,
+ punc_dict,
+ bpe_tokenizer,
+ self.dataset_conf,
+ self.frontend_conf,
+ speed_perturb=self.speed_perturb if mode == "train" else None,
+ mode=mode,
+ batch_mode=batch_mode,
+ )
def build_iter(self, epoch, shuffle=True):
self.dataset.set_epoch(epoch)
- data_loader = DataLoader(self.dataset,
- batch_size=None,
- pin_memory=True,
- num_workers=self.dataset_conf.get("num_workers", 8))
+ data_loader = DataLoader(
+ self.dataset,
+ batch_size=None,
+ pin_memory=True,
+ num_workers=self.dataset_conf.get("num_workers", 8),
+ )
return data_loader
--
Gitblit v1.9.1