OuroborosM commited on
Commit
9d0e9c8
·
1 Parent(s): fdba7fc

add tts model

Browse files
Files changed (1) hide show
  1. app.py +32 -0
app.py CHANGED
@@ -75,6 +75,38 @@ from langchain.schema import (
75
  LLMResult
76
  )
77
  import time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
78
 
79
  class GPTRemote(LLM):
80
  n: int
 
75
  LLMResult
76
  )
77
  import time
78
+ from datasets import load_dataset
79
+
80
+ from transformers import pipeline
81
+
82
+
83
+ ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
84
+ sample = ds[0]["audio"]
85
+
86
+ def speech_to_text_loc(audio):
87
+ device = "cpu"
88
+ pipe = pipeline(
89
+ "automatic-speech-recognition",
90
+ model="openai/whisper-small",
91
+ chunk_length_s=30,
92
+ device=device,
93
+ )
94
+ text = pipe(audio.copy(), batch_size=2)["text"]
95
+ return text
96
+
97
+ print("voice to text loc: ", speech_to_text_loc(sample))
98
+
99
+ def text_to_speech_loc(audio):
100
+ device = "cpu"
101
+ pipe = pipeline(
102
+ "text-to-speech",
103
+ model="microsoft/speecht5_tts",
104
+ device=device,
105
+ )
106
+ text = pipe(audio.copy(), batch_size=2)["text"]
107
+ return text
108
+
109
+ print("text to speech: ", text_to_speech_loc("Good morning."))
110
 
111
  class GPTRemote(LLM):
112
  n: int