Spaces:
Sleeping
Sleeping
Commit
·
7e182ab
1
Parent(s):
04716fa
Added textbox for unprocessed transcription string
Browse files
app.py
CHANGED
@@ -10,10 +10,6 @@ from transformers import (
|
|
10 |
)
|
11 |
|
12 |
|
13 |
-
# diction_text = """
|
14 |
-
# How is this leisure to be disposed of? In the public-house? the singing hall? the dancing-saloon? which hold out seductions somewhat more dangerous, methinks, to honest labor than those presented by a library; or in listless inaction, in weary unoccupied solitude? That cannot be. While man is a social animal society he must have, and better a thousand times that he should seek relief from the tedium of unemployed hours in the improving conversation of worthy authors, dead or living, than in the debasing, brutalising communications from which it is so difficult otherwise to escape.
|
15 |
-
# """
|
16 |
-
|
17 |
diction_text = """
|
18 |
How is this leisure to be disposed of? In the public-house? the singing hall? the dancing-saloon?
|
19 |
which hold out seductions somewhat more dangerous, methinks, to honest labour than those presented by a library...
|
@@ -26,7 +22,6 @@ device = "cpu"
|
|
26 |
|
27 |
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
|
28 |
|
29 |
-
# model_id = "openai/whisper-base "
|
30 |
model_id = "openai/whisper-large-v3"
|
31 |
|
32 |
description = f"""
|
@@ -87,17 +82,20 @@ def transcribe_audio(diction_text, audio):
|
|
87 |
|
88 |
formatted_score = f"{str(round(score,3))}%"
|
89 |
|
90 |
-
return diff_text, formatted_score
|
|
|
|
|
|
|
91 |
|
92 |
|
93 |
highlighted_results = gr.HighlightedText(
|
94 |
label="Diff",
|
95 |
combine_adjacent=True,
|
96 |
show_legend=True,
|
97 |
-
color_map={"+": "
|
98 |
)
|
99 |
|
100 |
-
score = gr.Textbox("0%", label="
|
101 |
|
102 |
|
103 |
input_audio = gr.Audio(
|
@@ -113,7 +111,7 @@ input_audio = gr.Audio(
|
|
113 |
|
114 |
demo = gr.Interface(
|
115 |
fn=transcribe_audio,
|
116 |
-
inputs=[diction_script, input_audio],
|
117 |
outputs=[highlighted_results, score],
|
118 |
title="Redmond Barry-oke",
|
119 |
description=description,
|
|
|
10 |
)
|
11 |
|
12 |
|
|
|
|
|
|
|
|
|
13 |
diction_text = """
|
14 |
How is this leisure to be disposed of? In the public-house? the singing hall? the dancing-saloon?
|
15 |
which hold out seductions somewhat more dangerous, methinks, to honest labour than those presented by a library...
|
|
|
22 |
|
23 |
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
|
24 |
|
|
|
25 |
model_id = "openai/whisper-large-v3"
|
26 |
|
27 |
description = f"""
|
|
|
82 |
|
83 |
formatted_score = f"{str(round(score,3))}%"
|
84 |
|
85 |
+
return diction_text, diff_text, formatted_score
|
86 |
+
|
87 |
+
|
88 |
+
transcribed_text = gr.Textbox(label="Transcribed text")
|
89 |
|
90 |
|
91 |
highlighted_results = gr.HighlightedText(
|
92 |
label="Diff",
|
93 |
combine_adjacent=True,
|
94 |
show_legend=True,
|
95 |
+
color_map={"+": "green", "-": "red"},
|
96 |
)
|
97 |
|
98 |
+
score = gr.Textbox("0%", label="Barry-oke score")
|
99 |
|
100 |
|
101 |
input_audio = gr.Audio(
|
|
|
111 |
|
112 |
demo = gr.Interface(
|
113 |
fn=transcribe_audio,
|
114 |
+
inputs=[transcribed_text, diction_script, input_audio],
|
115 |
outputs=[highlighted_results, score],
|
116 |
title="Redmond Barry-oke",
|
117 |
description=description,
|