Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -25,7 +25,7 @@ pipe = pipeline(
|
|
| 25 |
|
| 26 |
def transcribe(inputs, task):
|
| 27 |
if inputs is None:
|
| 28 |
-
raise gr.Error("
|
| 29 |
|
| 30 |
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
|
| 31 |
return text
|
|
@@ -41,29 +41,29 @@ def _return_yt_html_embed(yt_url):
|
|
| 41 |
|
| 42 |
def download_yt_audio(yt_url, filename):
|
| 43 |
info_loader = youtube_dl.YoutubeDL()
|
| 44 |
-
|
| 45 |
try:
|
| 46 |
info = info_loader.extract_info(yt_url, download=False)
|
| 47 |
except youtube_dl.utils.DownloadError as err:
|
| 48 |
raise gr.Error(str(err))
|
| 49 |
-
|
| 50 |
file_length = info["duration_string"]
|
| 51 |
file_h_m_s = file_length.split(":")
|
| 52 |
file_h_m_s = [int(sub_length) for sub_length in file_h_m_s]
|
| 53 |
-
|
| 54 |
if len(file_h_m_s) == 1:
|
| 55 |
file_h_m_s.insert(0, 0)
|
| 56 |
if len(file_h_m_s) == 2:
|
| 57 |
file_h_m_s.insert(0, 0)
|
| 58 |
file_length_s = file_h_m_s[0] * 3600 + file_h_m_s[1] * 60 + file_h_m_s[2]
|
| 59 |
-
|
| 60 |
if file_length_s > YT_LENGTH_LIMIT_S:
|
| 61 |
yt_length_limit_hms = time.strftime("%HH:%MM:%SS", time.gmtime(YT_LENGTH_LIMIT_S))
|
| 62 |
file_length_hms = time.strftime("%HH:%MM:%SS", time.gmtime(file_length_s))
|
| 63 |
-
raise gr.Error(f"
|
| 64 |
-
|
| 65 |
ydl_opts = {"outtmpl": filename, "format": "worstvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best"}
|
| 66 |
-
|
| 67 |
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
|
| 68 |
try:
|
| 69 |
ydl.download([yt_url])
|
|
@@ -99,40 +99,46 @@ mf_transcribe = gr.Interface(
|
|
| 99 |
outputs="text",
|
| 100 |
layout="horizontal",
|
| 101 |
theme="huggingface",
|
| 102 |
-
title="Whisper Large V3
|
| 103 |
-
description=
|
|
|
|
|
|
|
| 104 |
allow_flagging="never",
|
| 105 |
)
|
| 106 |
|
| 107 |
file_transcribe = gr.Interface(
|
| 108 |
fn=transcribe,
|
| 109 |
inputs=[
|
| 110 |
-
gr.inputs.Audio(source="upload", type="filepath", optional=True, label="Audio
|
| 111 |
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
|
| 112 |
],
|
| 113 |
outputs="text",
|
| 114 |
layout="horizontal",
|
| 115 |
theme="huggingface",
|
| 116 |
-
title="Whisper Large V3
|
| 117 |
-
description=
|
|
|
|
|
|
|
| 118 |
allow_flagging="never",
|
| 119 |
)
|
| 120 |
|
| 121 |
yt_transcribe = gr.Interface(
|
| 122 |
fn=yt_transcribe,
|
| 123 |
inputs=[
|
| 124 |
-
gr.inputs.Textbox(lines=1, placeholder="
|
| 125 |
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe")
|
| 126 |
],
|
| 127 |
outputs=["html", "text"],
|
| 128 |
layout="horizontal",
|
| 129 |
theme="huggingface",
|
| 130 |
-
title="Whisper Large V3
|
| 131 |
-
description=
|
|
|
|
|
|
|
| 132 |
allow_flagging="never",
|
| 133 |
)
|
| 134 |
|
| 135 |
with demo:
|
| 136 |
-
gr.TabbedInterface([mf_transcribe, file_transcribe, yt_transcribe], ["
|
| 137 |
|
| 138 |
demo.launch(enable_queue=True)
|
|
|
|
| 25 |
|
| 26 |
def transcribe(inputs, task):
|
| 27 |
if inputs is None:
|
| 28 |
+
raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
|
| 29 |
|
| 30 |
text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
|
| 31 |
return text
|
|
|
|
| 41 |
|
| 42 |
def download_yt_audio(yt_url, filename):
|
| 43 |
info_loader = youtube_dl.YoutubeDL()
|
| 44 |
+
|
| 45 |
try:
|
| 46 |
info = info_loader.extract_info(yt_url, download=False)
|
| 47 |
except youtube_dl.utils.DownloadError as err:
|
| 48 |
raise gr.Error(str(err))
|
| 49 |
+
|
| 50 |
file_length = info["duration_string"]
|
| 51 |
file_h_m_s = file_length.split(":")
|
| 52 |
file_h_m_s = [int(sub_length) for sub_length in file_h_m_s]
|
| 53 |
+
|
| 54 |
if len(file_h_m_s) == 1:
|
| 55 |
file_h_m_s.insert(0, 0)
|
| 56 |
if len(file_h_m_s) == 2:
|
| 57 |
file_h_m_s.insert(0, 0)
|
| 58 |
file_length_s = file_h_m_s[0] * 3600 + file_h_m_s[1] * 60 + file_h_m_s[2]
|
| 59 |
+
|
| 60 |
if file_length_s > YT_LENGTH_LIMIT_S:
|
| 61 |
yt_length_limit_hms = time.strftime("%HH:%MM:%SS", time.gmtime(YT_LENGTH_LIMIT_S))
|
| 62 |
file_length_hms = time.strftime("%HH:%MM:%SS", time.gmtime(file_length_s))
|
| 63 |
+
raise gr.Error(f"Maximum YouTube length is {yt_length_limit_hms}, got {file_length_hms} YouTube video.")
|
| 64 |
+
|
| 65 |
ydl_opts = {"outtmpl": filename, "format": "worstvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best"}
|
| 66 |
+
|
| 67 |
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
|
| 68 |
try:
|
| 69 |
ydl.download([yt_url])
|
|
|
|
| 99 |
outputs="text",
|
| 100 |
layout="horizontal",
|
| 101 |
theme="huggingface",
|
| 102 |
+
title="Whisper Large V3: Transcribe Audio",
|
| 103 |
+
description=(
|
| 104 |
+
"Whisper Large V3 fine-tuned for Uzbek language by Dataprizma"
|
| 105 |
+
),
|
| 106 |
allow_flagging="never",
|
| 107 |
)
|
| 108 |
|
| 109 |
file_transcribe = gr.Interface(
|
| 110 |
fn=transcribe,
|
| 111 |
inputs=[
|
| 112 |
+
gr.inputs.Audio(source="upload", type="filepath", optional=True, label="Audio file"),
|
| 113 |
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
|
| 114 |
],
|
| 115 |
outputs="text",
|
| 116 |
layout="horizontal",
|
| 117 |
theme="huggingface",
|
| 118 |
+
title="Whisper Large V3: Transcribe Audio",
|
| 119 |
+
description=(
|
| 120 |
+
"Whisper Large V3 fine-tuned for Uzbek language by Dataprizma"
|
| 121 |
+
),
|
| 122 |
allow_flagging="never",
|
| 123 |
)
|
| 124 |
|
| 125 |
yt_transcribe = gr.Interface(
|
| 126 |
fn=yt_transcribe,
|
| 127 |
inputs=[
|
| 128 |
+
gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
|
| 129 |
gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe")
|
| 130 |
],
|
| 131 |
outputs=["html", "text"],
|
| 132 |
layout="horizontal",
|
| 133 |
theme="huggingface",
|
| 134 |
+
title="Whisper Large V3: Transcribe YouTube",
|
| 135 |
+
description=(
|
| 136 |
+
"Whisper Large V3 fine-tuned for Uzbek language by Dataprizma"
|
| 137 |
+
),
|
| 138 |
allow_flagging="never",
|
| 139 |
)
|
| 140 |
|
| 141 |
with demo:
|
| 142 |
+
gr.TabbedInterface([mf_transcribe, file_transcribe, yt_transcribe], ["Microphone", "Audio file", "YouTube"])
|
| 143 |
|
| 144 |
demo.launch(enable_queue=True)
|