From 4ba1011b42e041ee1d71448eefd7ef2e7bd61bb6 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期五, 31 三月 2023 15:31:26 +0800
Subject: [PATCH] export
---
funasr/bin/asr_inference_uniasr.py | 33 ++++++++++++++++-----------------
1 files changed, 16 insertions(+), 17 deletions(-)
diff --git a/funasr/bin/asr_inference_uniasr.py b/funasr/bin/asr_inference_uniasr.py
index c50bf17..1286bc2 100644
--- a/funasr/bin/asr_inference_uniasr.py
+++ b/funasr/bin/asr_inference_uniasr.py
@@ -37,16 +37,13 @@
from funasr.models.frontend.wav_frontend import WavFrontend
-header_colors = '\033[95m'
-end_colors = '\033[0m'
-
class Speech2Text:
"""Speech2Text class
Examples:
>>> import soundfile
- >>> speech2text = Speech2Text("asr_config.yml", "asr.pth")
+ >>> speech2text = Speech2Text("asr_config.yml", "asr.pb")
>>> audio, rate = soundfile.read("speech.wav")
>>> speech2text(audio)
[(text, token, token_int, hypothesis object), ...]
@@ -261,6 +258,7 @@
# Change integer-ids to tokens
token = self.converter.ids2tokens(token_int)
+ token = list(filter(lambda x: x != "<gbg>", token))
if self.tokenizer is not None:
text = self.tokenizer.tokens2text(token)
@@ -398,6 +396,19 @@
else:
device = "cpu"
+ if param_dict is not None and "decoding_model" in param_dict:
+ if param_dict["decoding_model"] == "fast":
+ decoding_ind = 0
+ decoding_mode = "model1"
+ elif param_dict["decoding_model"] == "normal":
+ decoding_ind = 0
+ decoding_mode = "model2"
+ elif param_dict["decoding_model"] == "offline":
+ decoding_ind = 1
+ decoding_mode = "model2"
+ else:
+ raise NotImplementedError("unsupported decoding model {}".format(param_dict["decoding_model"]))
+
# 1. Set random-seed
set_all_random_seed(seed)
@@ -440,18 +451,6 @@
if isinstance(raw_inputs, torch.Tensor):
raw_inputs = raw_inputs.numpy()
data_path_and_name_and_type = [raw_inputs, "speech", "waveform"]
- if param_dict is not None and "decoding_model" in param_dict:
- if param_dict["decoding_model"] == "fast":
- speech2text.decoding_ind = 0
- speech2text.decoding_mode = "model1"
- elif param_dict["decoding_model"] == "normal":
- speech2text.decoding_ind = 0
- speech2text.decoding_mode = "model2"
- elif param_dict["decoding_model"] == "offline":
- speech2text.decoding_ind = 1
- speech2text.decoding_mode = "model2"
- else:
- raise NotImplementedError("unsupported decoding model {}".format(param_dict["decoding_model"]))
loader = ASRTask.build_streaming_iterator(
data_path_and_name_and_type,
dtype=dtype,
@@ -511,7 +510,7 @@
finish_count += 1
asr_utils.print_progress(finish_count / file_count)
if writer is not None:
- ibest_writer["text"][key] = text
+ ibest_writer["text"][key] = text_postprocessed
return asr_result_list
return _forward
--
Gitblit v1.9.1