Spaces:
Running
on
T4
Running
on
T4
Update app.py
Browse files
app.py
CHANGED
|
@@ -21,18 +21,8 @@ transcribe_token_id = all_special_ids[-5]
|
|
| 21 |
translate_token_id = all_special_ids[-6]
|
| 22 |
|
| 23 |
|
| 24 |
-
def transcribe(microphone,
|
| 25 |
-
|
| 26 |
-
if (microphone is not None) and (file_upload is not None):
|
| 27 |
-
warn_output = (
|
| 28 |
-
"WARNING: You've uploaded an audio file and used the microphone. "
|
| 29 |
-
"The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
|
| 30 |
-
)
|
| 31 |
-
|
| 32 |
-
elif (microphone is None) and (file_upload is None):
|
| 33 |
-
return "ERROR: You have to either use the microphone or upload an audio file"
|
| 34 |
-
|
| 35 |
-
file = microphone if microphone is not None else file_upload
|
| 36 |
|
| 37 |
pipe.model.config.forced_decoder_ids = [[2, transcribe_token_id if task=="transcribe" else translate_token_id]]
|
| 38 |
|
|
@@ -68,9 +58,8 @@ demo = gr.Blocks()
|
|
| 68 |
mf_transcribe = gr.Interface(
|
| 69 |
fn=transcribe,
|
| 70 |
inputs=[
|
| 71 |
-
gr.
|
| 72 |
-
gr.
|
| 73 |
-
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
|
| 74 |
],
|
| 75 |
outputs="text",
|
| 76 |
layout="horizontal",
|
|
@@ -84,26 +73,8 @@ mf_transcribe = gr.Interface(
|
|
| 84 |
allow_flagging="never",
|
| 85 |
)
|
| 86 |
|
| 87 |
-
yt_transcribe = gr.Interface(
|
| 88 |
-
fn=yt_transcribe,
|
| 89 |
-
inputs=[
|
| 90 |
-
gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
|
| 91 |
-
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe")
|
| 92 |
-
],
|
| 93 |
-
outputs=["html", "text"],
|
| 94 |
-
layout="horizontal",
|
| 95 |
-
theme="huggingface",
|
| 96 |
-
title="Whisper Large V2: Transcribe YouTube",
|
| 97 |
-
description=(
|
| 98 |
-
"Transcribe long-form YouTube videos with the click of a button! Demo uses the checkpoint"
|
| 99 |
-
f" [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe video files of"
|
| 100 |
-
" arbitrary length."
|
| 101 |
-
),
|
| 102 |
-
allow_flagging="never",
|
| 103 |
-
)
|
| 104 |
|
| 105 |
-
|
| 106 |
-
gr.TabbedInterface([mf_transcribe, yt_transcribe], ["Transcribe Audio", "Transcribe YouTube"])
|
| 107 |
|
| 108 |
demo.launch(enable_queue=True)
|
| 109 |
|
|
|
|
| 21 |
translate_token_id = all_special_ids[-6]
|
| 22 |
|
| 23 |
|
| 24 |
+
def transcribe(microphone, task):
|
| 25 |
+
file = microphone
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
pipe.model.config.forced_decoder_ids = [[2, transcribe_token_id if task=="transcribe" else translate_token_id]]
|
| 28 |
|
|
|
|
| 58 |
mf_transcribe = gr.Interface(
|
| 59 |
fn=transcribe,
|
| 60 |
inputs=[
|
| 61 |
+
gr.Audio(source="microphone", type="filepath", optional=True),
|
| 62 |
+
gr.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
|
|
|
|
| 63 |
],
|
| 64 |
outputs="text",
|
| 65 |
layout="horizontal",
|
|
|
|
| 73 |
allow_flagging="never",
|
| 74 |
)
|
| 75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
|
| 77 |
+
|
|
|
|
| 78 |
|
| 79 |
demo.launch(enable_queue=True)
|
| 80 |
|