From bdfd27b9e96bd55c449953bb577e1d4deeaf11c9 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期六, 13 一月 2024 23:43:17 +0800
Subject: [PATCH] funasr1.0
---
funasr/utils/load_utils.py | 32 +++++++++++++++++++++-----------
1 files changed, 21 insertions(+), 11 deletions(-)
diff --git a/funasr/utils/load_utils.py b/funasr/utils/load_utils.py
index c82987f..4e131a8 100644
--- a/funasr/utils/load_utils.py
+++ b/funasr/utils/load_utils.py
@@ -16,7 +16,7 @@
-def load_audio_text_image_video(data_or_path_or_list, fs: int = 16000, audio_fs: int = 16000, data_type="sound", tokenizer=None):
+def load_audio_text_image_video(data_or_path_or_list, fs: int = 16000, audio_fs: int = 16000, data_type="sound", tokenizer=None, **kwargs):
if isinstance(data_or_path_or_list, (list, tuple)):
if data_type is not None and isinstance(data_type, (list, tuple)):
@@ -26,24 +26,34 @@
for j, (data_type_j, data_or_path_or_list_j) in enumerate(zip(data_type_i, data_or_path_or_list_i)):
- data_or_path_or_list_j = load_audio_text_image_video(data_or_path_or_list_j, fs=fs, audio_fs=audio_fs, data_type=data_type_j, tokenizer=tokenizer)
+ data_or_path_or_list_j = load_audio_text_image_video(data_or_path_or_list_j, fs=fs, audio_fs=audio_fs, data_type=data_type_j, tokenizer=tokenizer, **kwargs)
data_or_path_or_list_ret[j].append(data_or_path_or_list_j)
return data_or_path_or_list_ret
else:
- return [load_audio_text_image_video(audio, fs=fs, audio_fs=audio_fs, data_type=data_type) for audio in data_or_path_or_list]
- if isinstance(data_or_path_or_list, str) and data_or_path_or_list.startswith('http'):
+ return [load_audio_text_image_video(audio, fs=fs, audio_fs=audio_fs, data_type=data_type, **kwargs) for audio in data_or_path_or_list]
+
+ if isinstance(data_or_path_or_list, str) and data_or_path_or_list.startswith('http'): # download url to local file
data_or_path_or_list = download_from_url(data_or_path_or_list)
- if isinstance(data_or_path_or_list, str) and os.path.exists(data_or_path_or_list):
+
+ if isinstance(data_or_path_or_list, str) and os.path.exists(data_or_path_or_list): # local file
if data_type is None or data_type == "sound":
data_or_path_or_list, audio_fs = torchaudio.load(data_or_path_or_list)
data_or_path_or_list = data_or_path_or_list[0, :]
- # elif data_type == "text" and tokenizer is not None:
- # data_or_path_or_list = tokenizer.encode(data_or_path_or_list)
+ elif data_type == "text" and tokenizer is not None:
+ data_or_path_or_list = tokenizer.encode(data_or_path_or_list)
+ elif data_type == "image": # undo
+ pass
+ elif data_type == "video": # undo
+ pass
+
+ # if data_in is a file or url, set is_final=True
+ if "cache" in kwargs:
+ kwargs["cache"]["is_final"] = True
elif isinstance(data_or_path_or_list, str) and data_type == "text" and tokenizer is not None:
data_or_path_or_list = tokenizer.encode(data_or_path_or_list)
elif isinstance(data_or_path_or_list, np.ndarray): # audio sample point
- data_or_path_or_list = np.squeeze(data_or_path_or_list) # [n_samples,]
+ data_or_path_or_list = torch.from_numpy(data_or_path_or_list).squeeze() # [n_samples,]
else:
pass
# print(f"unsupport data type: {data_or_path_or_list}, return raw data")
@@ -68,7 +78,7 @@
array = np.frombuffer((middle_data.astype(dtype) - offset) / abs_max, dtype=np.float32)
return array
-def extract_fbank(data, data_len = None, data_type: str="sound", frontend=None):
+def extract_fbank(data, data_len = None, data_type: str="sound", frontend=None, **kwargs):
# import pdb;
# pdb.set_trace()
if isinstance(data, np.ndarray):
@@ -83,7 +93,7 @@
elif isinstance(data, (list, tuple)):
data_list, data_len = [], []
for data_i in data:
- if isinstance(data, np.ndarray):
+ if isinstance(data_i, np.ndarray):
data_i = torch.from_numpy(data_i)
data_list.append(data_i)
data_len.append(data_i.shape[0])
@@ -91,7 +101,7 @@
# import pdb;
# pdb.set_trace()
# if data_type == "sound":
- data, data_len = frontend(data, data_len)
+ data, data_len = frontend(data, data_len, **kwargs)
if isinstance(data_len, (list, tuple)):
data_len = torch.tensor([data_len])
--
Gitblit v1.9.1