jthteo commited on
Commit
d089c2e
1 Parent(s): 9dc7dd3

Added choice for Model Size

Browse files
Files changed (1) hide show
  1. app.py +17 -3
app.py CHANGED
@@ -4,6 +4,13 @@ import gradio as gr
4
  import whisper
5
 
6
  model = whisper.load_model("small")
 
 
 
 
 
 
 
7
 
8
 
9
  def inference(audio):
@@ -131,13 +138,17 @@ with block:
131
  <p style="margin-bottom: 10px; font-size: 94%">
132
  Whisper is a general-purpose speech recognition model. It has been trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech recognition as well as speech translation and language identification. </p>
133
  <p>This is a fork by JTHTEO.</p>
134
- <p>This uses the small multi-lingual 244mill parameter Whisper model (<a href="https://github.com/openai/whisper/blob/main/model-card.md">Model Card</a>.) </p>
135
  </p>
136
  </div>
137
  """
138
  )
139
  with gr.Group():
140
  with gr.Box():
 
 
 
 
141
  with gr.Row().style(mobile_collapse=False, equal_height=True):
142
  audio = gr.Audio(
143
  label="Input Audio",
@@ -147,10 +158,13 @@ with block:
147
  )
148
  btn = gr.Button("Transcribe")
149
  text = gr.Textbox(show_label=False)
150
-
151
-
 
152
  btn.click(inference, inputs=[audio], outputs=[text])
153
 
 
 
154
  gr.HTML('''
155
  <div class="footer">
156
  <p>Model by <a href="https://github.com/openai/whisper" style="text-decoration: underline;" target="_blank">OpenAI</a> - Gradio Demo by 🤗 Hugging Face, this is a fork by JTHTEO
 
4
  import whisper
5
 
6
  model = whisper.load_model("small")
7
+ current_size = 'small'
8
+
9
+ def change_model(size):
10
+ if size == current_size:
11
+ return
12
+ model = whisper.load_model(size)
13
+ current_size = size
14
 
15
 
16
  def inference(audio):
 
138
  <p style="margin-bottom: 10px; font-size: 94%">
139
  Whisper is a general-purpose speech recognition model. It has been trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech recognition as well as speech translation and language identification. </p>
140
  <p>This is a fork by JTHTEO.</p>
141
+ <p>The different sized Whisper models can be found in this (<a href="https://github.com/openai/whisper/blob/main/model-card.md">Model Card</a>.) </p>
142
  </p>
143
  </div>
144
  """
145
  )
146
  with gr.Group():
147
  with gr.Box():
148
+ wmodel = gr.Radio(
149
+ choices=["tiny", "base", "small", "medium", "large"],
150
+ label="Model used",
151
+ value="small")
152
  with gr.Row().style(mobile_collapse=False, equal_height=True):
153
  audio = gr.Audio(
154
  label="Input Audio",
 
158
  )
159
  btn = gr.Button("Transcribe")
160
  text = gr.Textbox(show_label=False)
161
+
162
+ ##events###
163
+ wmodel.change(change_model, inputs=[wmodel], outputs=[])
164
  btn.click(inference, inputs=[audio], outputs=[text])
165
 
166
+ ##footer###
167
+
168
  gr.HTML('''
169
  <div class="footer">
170
  <p>Model by <a href="https://github.com/openai/whisper" style="text-decoration: underline;" target="_blank">OpenAI</a> - Gradio Demo by 🤗 Hugging Face, this is a fork by JTHTEO