Update app.py
Browse files
app.py
CHANGED
|
@@ -36,7 +36,11 @@ def diffuse(prompt, guidance_scale, nb_iterations, seed):
|
|
| 36 |
|
| 37 |
generator = torch.Generator(device=device).manual_seed(int(seed))
|
| 38 |
|
| 39 |
-
print("
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
print("prompt: " + prompt)
|
| 41 |
print("guidance scale: " + str(guidance_scale))
|
| 42 |
print("inference steps: " + str(nb_iterations))
|
|
@@ -63,6 +67,11 @@ def diffuse(prompt, guidance_scale, nb_iterations, seed):
|
|
| 63 |
return images
|
| 64 |
|
| 65 |
def translate(audio):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
audio = whisper.load_audio(audio)
|
| 67 |
audio = whisper.pad_or_trim(audio)
|
| 68 |
|
|
@@ -77,8 +86,8 @@ def translate(audio):
|
|
| 77 |
translation = whisper.decode(whisper_model, mel, translate_options)
|
| 78 |
|
| 79 |
print("language spoken: " + transcription.language)
|
| 80 |
-
print(transcription.text)
|
| 81 |
-
print(translation.text)
|
| 82 |
|
| 83 |
return transcription.language, transcription.text, translation.text
|
| 84 |
|
|
@@ -144,8 +153,11 @@ css = """
|
|
| 144 |
padding: 20px;
|
| 145 |
height: 160px;
|
| 146 |
}
|
| 147 |
-
#
|
| 148 |
-
|
|
|
|
|
|
|
|
|
|
| 149 |
}
|
| 150 |
div#spoken_lang textarea {
|
| 151 |
font-size: 4em;
|
|
@@ -197,7 +209,7 @@ with gr.Blocks(css=css) as demo:
|
|
| 197 |
</p>
|
| 198 |
|
| 199 |
<p style='text-align: center;'>
|
| 200 |
-
This demo is running on π’ CPU
|
| 201 |
β
|
| 202 |
</p>
|
| 203 |
|
|
@@ -222,14 +234,19 @@ with gr.Blocks(css=css) as demo:
|
|
| 222 |
)
|
| 223 |
with gr.Row():
|
| 224 |
audio_r_translate = gr.Button("Check Whisper first ? π")
|
| 225 |
-
audio_r_direct_sd = gr.Button("Magic Whisper
|
| 226 |
|
| 227 |
with gr.Tab(label="Upload audio input", elem_id="upload_tab"):
|
| 228 |
with gr.Column():
|
| 229 |
-
upload_input = gr.Audio(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 230 |
with gr.Row():
|
| 231 |
-
audio_u_translate = gr.Button("Check Whisper first ?")
|
| 232 |
-
audio_u_direct_sd = gr.Button("Magic Whisper
|
| 233 |
|
| 234 |
with gr.Accordion(label="Stable Diffusion Settings", elem_id="sd_settings"):
|
| 235 |
with gr.Row():
|
|
|
|
| 36 |
|
| 37 |
generator = torch.Generator(device=device).manual_seed(int(seed))
|
| 38 |
|
| 39 |
+
print("""
|
| 40 |
+
β
|
| 41 |
+
Sending prompt to Stable Diffusion ...
|
| 42 |
+
β
|
| 43 |
+
""")
|
| 44 |
print("prompt: " + prompt)
|
| 45 |
print("guidance scale: " + str(guidance_scale))
|
| 46 |
print("inference steps: " + str(nb_iterations))
|
|
|
|
| 67 |
return images
|
| 68 |
|
| 69 |
def translate(audio):
|
| 70 |
+
print("""
|
| 71 |
+
β
|
| 72 |
+
Sending audio to Whisper ...
|
| 73 |
+
β
|
| 74 |
+
""")
|
| 75 |
audio = whisper.load_audio(audio)
|
| 76 |
audio = whisper.pad_or_trim(audio)
|
| 77 |
|
|
|
|
| 86 |
translation = whisper.decode(whisper_model, mel, translate_options)
|
| 87 |
|
| 88 |
print("language spoken: " + transcription.language)
|
| 89 |
+
print("transcript: " + transcription.text)
|
| 90 |
+
print("translated: " + translation.text)
|
| 91 |
|
| 92 |
return transcription.language, transcription.text, translation.text
|
| 93 |
|
|
|
|
| 153 |
padding: 20px;
|
| 154 |
height: 160px;
|
| 155 |
}
|
| 156 |
+
div#upload_area {
|
| 157 |
+
height: 11.1rem;
|
| 158 |
+
}
|
| 159 |
+
div#upload_area > div.w-full > div {
|
| 160 |
+
min-height: 9rem;
|
| 161 |
}
|
| 162 |
div#spoken_lang textarea {
|
| 163 |
font-size: 4em;
|
|
|
|
| 209 |
</p>
|
| 210 |
|
| 211 |
<p style='text-align: center;'>
|
| 212 |
+
This demo is running on π’ CPU β’ Offered by Sylvain <a href='https://twitter.com/fffiloni' target='_blank'>@fffiloni</a> β’ <img id='visitor-badge' alt='visitor badge' src='https://visitor-badge.glitch.me/badge?page_id=gradio-blocks.whisper-to-stable-diffusion' style='display: inline-block' /><br />
|
| 213 |
β
|
| 214 |
</p>
|
| 215 |
|
|
|
|
| 234 |
)
|
| 235 |
with gr.Row():
|
| 236 |
audio_r_translate = gr.Button("Check Whisper first ? π")
|
| 237 |
+
audio_r_direct_sd = gr.Button("Magic Whisper βΊ SD right now! π€ ")
|
| 238 |
|
| 239 |
with gr.Tab(label="Upload audio input", elem_id="upload_tab"):
|
| 240 |
with gr.Column():
|
| 241 |
+
upload_input = gr.Audio(
|
| 242 |
+
source="upload",
|
| 243 |
+
type="filepath",
|
| 244 |
+
show_label=False,
|
| 245 |
+
elem_id="upload_area"
|
| 246 |
+
)
|
| 247 |
with gr.Row():
|
| 248 |
+
audio_u_translate = gr.Button("Check Whisper first ? π")
|
| 249 |
+
audio_u_direct_sd = gr.Button("Magic Whisper βΊ SD right now! π€ ")
|
| 250 |
|
| 251 |
with gr.Accordion(label="Stable Diffusion Settings", elem_id="sd_settings"):
|
| 252 |
with gr.Row():
|