Mark Duppenthaler commited on
Commit
7fb1760
1 Parent(s): 366edf8

dummpy outputs

Browse files
__pycache__/sample_wav.cpython-310.pyc ADDED
Binary file (136 kB). View file
 
__pycache__/simuleval_transcoder.cpython-310.pyc CHANGED
Binary files a/__pycache__/simuleval_transcoder.cpython-310.pyc and b/__pycache__/simuleval_transcoder.cpython-310.pyc differ
 
app.py CHANGED
@@ -6,6 +6,9 @@ import gradio as gr
6
  import numpy as np
7
  import torch
8
  import torchaudio
 
 
 
9
 
10
  from simuleval_transcoder import *
11
 
@@ -45,6 +48,15 @@ def translate_audio_segment(audio):
45
  logger.debug(f"translate_audio_segment: incoming audio")
46
  sample_rate, data = audio
47
 
 
 
 
 
 
 
 
 
 
48
  transcoder.process_incoming_bytes(data.tobytes(), 'eng', sample_rate)
49
 
50
  speech_and_text_output = transcoder.get_buffered_output()
@@ -67,6 +79,9 @@ def translate_audio_segment(audio):
67
 
68
  return speech, text
69
 
 
 
 
70
  def streaming_input_callback(
71
  audio_file, translated_audio_bytes_state, translated_text_state
72
  ):
@@ -76,9 +91,14 @@ def streaming_input_callback(
76
 
77
  # TODO: accumulate each segment to provide a continuous audio segment
78
 
 
 
 
79
  if translated_wav_segment is not None:
80
  sample_rate, audio_bytes = translated_wav_segment
81
- audio_np_array = np.frombuffer(audio_bytes, dtype=np.float32, count=3)
 
 
82
 
83
 
84
  # combine translated wav
 
6
  import numpy as np
7
  import torch
8
  import torchaudio
9
+ import sys
10
+ from sample_wav import sample_wav
11
+ np.set_printoptions(threshold=sys.maxsize)
12
 
13
  from simuleval_transcoder import *
14
 
 
48
  logger.debug(f"translate_audio_segment: incoming audio")
49
  sample_rate, data = audio
50
 
51
+ # print(sample_rate)
52
+ # print("--------- start \n")
53
+ # # print(data)
54
+ # def map(x):
55
+ # return x
56
+ # print(data.tolist())
57
+ # print("--------- end \n")
58
+
59
+
60
  transcoder.process_incoming_bytes(data.tobytes(), 'eng', sample_rate)
61
 
62
  speech_and_text_output = transcoder.get_buffered_output()
 
79
 
80
  return speech, text
81
 
82
+ def dummy_ouput():
83
+ np.array()
84
+
85
  def streaming_input_callback(
86
  audio_file, translated_audio_bytes_state, translated_text_state
87
  ):
 
91
 
92
  # TODO: accumulate each segment to provide a continuous audio segment
93
 
94
+ # TEMP
95
+ translated_wav_segment = (46_000, sample_wav())
96
+
97
  if translated_wav_segment is not None:
98
  sample_rate, audio_bytes = translated_wav_segment
99
+ # TODO: convert to 16 bit int
100
+ # audio_np_array = np.frombuffer(audio_bytes, dtype=np.float32, count=3)
101
+ audio_np_array = audio_bytes
102
 
103
 
104
  # combine translated wav
sample_wav.py ADDED
The diff for this file is too large to render. See raw diff