From 74aa12ee4bbef787236bd382b186a17db40866a6 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期一, 18 九月 2023 10:43:27 +0800
Subject: [PATCH] docs
---
funasr/bin/asr_inference_launch.py | 12 +++++++-----
1 files changed, 7 insertions(+), 5 deletions(-)
diff --git a/funasr/bin/asr_inference_launch.py b/funasr/bin/asr_inference_launch.py
index cdaaefc..ea0f221 100644
--- a/funasr/bin/asr_inference_launch.py
+++ b/funasr/bin/asr_inference_launch.py
@@ -45,7 +45,7 @@
from funasr.utils.types import str2triple_str
from funasr.utils.types import str_or_none
from funasr.utils.vad_utils import slice_padding_fbank
-
+from tqdm import tqdm
def inference_asr(
maxlenratio: float,
@@ -651,7 +651,8 @@
batch_size_token_ms_cum = 0
beg_idx = 0
- for j, _ in enumerate(range(0, n)):
+ beg_asr_total = time.time()
+ for j, _ in tqdm(enumerate(range(0, n))):
batch_size_token_ms_cum += (sorted_data[j][0][1] - sorted_data[j][0][0])
if j < n - 1 and (batch_size_token_ms_cum + sorted_data[j + 1][0][1] - sorted_data[j + 1][0][0]) < batch_size_token_ms and (sorted_data[j + 1][0][1] - sorted_data[j + 1][0][0]) < batch_size_token_threshold_s:
continue
@@ -661,16 +662,17 @@
beg_idx = end_idx
batch = {"speech": speech_j, "speech_lengths": speech_lengths_j}
batch = to_device(batch, device=device)
- print("batch: ", speech_j.shape[0])
+ # print("batch: ", speech_j.shape[0])
beg_asr = time.time()
results = speech2text(**batch)
end_asr = time.time()
- print("time cost asr: ", end_asr - beg_asr)
+ # print("time cost asr: ", end_asr - beg_asr)
if len(results) < 1:
results = [["", [], [], [], [], [], []]]
results_sorted.extend(results)
-
+ end_asr_total = time.time()
+ print("total time cost asr: ", end_asr_total-beg_asr_total)
restored_data = [0] * n
for j in range(n):
index = sorted_data[j][1]
--
Gitblit v1.9.1