csukuangfj
commited on
Commit
·
81687e9
1
Parent(s):
f5b2e32
Minor fixes
Browse files
app.py
CHANGED
@@ -49,6 +49,9 @@ def process(
|
|
49 |
print("in_filename", in_filename)
|
50 |
print("language", language)
|
51 |
print("repo_id", repo_id)
|
|
|
|
|
|
|
52 |
filename = convert_to_wav(in_filename)
|
53 |
|
54 |
now = datetime.now()
|
@@ -89,7 +92,16 @@ def process(
|
|
89 |
print("hyp")
|
90 |
print(hyp)
|
91 |
|
92 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93 |
|
94 |
|
95 |
title = "# Automatic Speech Recognition with Next-gen Kaldi"
|
@@ -113,7 +125,16 @@ def update_model_dropdown(language: str):
|
|
113 |
raise ValueError(f"Unsupported language: {language}")
|
114 |
|
115 |
|
116 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
117 |
|
118 |
with demo:
|
119 |
gr.Markdown(title)
|
@@ -124,7 +145,11 @@ with demo:
|
|
124 |
choices=language_choices,
|
125 |
value=language_choices[0],
|
126 |
)
|
127 |
-
model_dropdown = gr.Dropdown(
|
|
|
|
|
|
|
|
|
128 |
language_radio.change(
|
129 |
update_model_dropdown,
|
130 |
inputs=language_radio,
|
@@ -146,29 +171,25 @@ with demo:
|
|
146 |
|
147 |
with gr.Tabs():
|
148 |
with gr.TabItem("Upload from disk"):
|
149 |
-
uploaded_file = gr.
|
150 |
source="upload", # Choose between "microphone", "upload"
|
151 |
type="filepath",
|
152 |
optional=False,
|
153 |
label="Upload from disk",
|
154 |
)
|
|
|
155 |
upload_button = gr.Button("Submit for recognition")
|
156 |
-
uploaded_output = gr.outputs.Textbox(
|
157 |
-
label="Recognized speech from uploaded file"
|
158 |
-
)
|
159 |
|
160 |
with gr.TabItem("Record from microphone"):
|
161 |
-
microphone = gr.
|
162 |
source="microphone", # Choose between "microphone", "upload"
|
163 |
type="filepath",
|
164 |
optional=False,
|
165 |
label="Record from microphone",
|
166 |
)
|
167 |
-
recorded_output = gr.outputs.Textbox(
|
168 |
-
label="Recognized speech from recordings"
|
169 |
-
)
|
170 |
|
171 |
record_button = gr.Button("Submit for recognition")
|
|
|
172 |
|
173 |
upload_button.click(
|
174 |
process,
|
|
|
49 |
print("in_filename", in_filename)
|
50 |
print("language", language)
|
51 |
print("repo_id", repo_id)
|
52 |
+
print("decoding_method", decoding_method)
|
53 |
+
print("num_active_paths", num_active_paths)
|
54 |
+
|
55 |
filename = convert_to_wav(in_filename)
|
56 |
|
57 |
now = datetime.now()
|
|
|
92 |
print("hyp")
|
93 |
print(hyp)
|
94 |
|
95 |
+
html_output = f"""
|
96 |
+
<div class='result'>
|
97 |
+
<div class='result_item result_item_success'>
|
98 |
+
{hyp}
|
99 |
+
<br/>
|
100 |
+
</div>
|
101 |
+
</div>
|
102 |
+
"""
|
103 |
+
|
104 |
+
return html_output
|
105 |
|
106 |
|
107 |
title = "# Automatic Speech Recognition with Next-gen Kaldi"
|
|
|
125 |
raise ValueError(f"Unsupported language: {language}")
|
126 |
|
127 |
|
128 |
+
# The css style is copied from
|
129 |
+
# https://huggingface.co/spaces/alphacep/asr/blob/main/app.py#L112
|
130 |
+
demo = gr.Blocks(
|
131 |
+
css="""
|
132 |
+
.result {display:flex;flex-direction:column}
|
133 |
+
.result_item {padding:15px;margin-bottom:8px;border-radius:15px;width:100%}
|
134 |
+
.result_item_success {background-color:mediumaquamarine;color:white;align-self:start}
|
135 |
+
.result_item_error {background-color:#ff7070;color:white;align-self:start}
|
136 |
+
""",
|
137 |
+
)
|
138 |
|
139 |
with demo:
|
140 |
gr.Markdown(title)
|
|
|
145 |
choices=language_choices,
|
146 |
value=language_choices[0],
|
147 |
)
|
148 |
+
model_dropdown = gr.Dropdown(
|
149 |
+
choices=language_to_models[language_choices[0]],
|
150 |
+
label="Select a model",
|
151 |
+
)
|
152 |
+
|
153 |
language_radio.change(
|
154 |
update_model_dropdown,
|
155 |
inputs=language_radio,
|
|
|
171 |
|
172 |
with gr.Tabs():
|
173 |
with gr.TabItem("Upload from disk"):
|
174 |
+
uploaded_file = gr.Audio(
|
175 |
source="upload", # Choose between "microphone", "upload"
|
176 |
type="filepath",
|
177 |
optional=False,
|
178 |
label="Upload from disk",
|
179 |
)
|
180 |
+
uploaded_output = gr.HTML(label="Recognized speech from uploaded file")
|
181 |
upload_button = gr.Button("Submit for recognition")
|
|
|
|
|
|
|
182 |
|
183 |
with gr.TabItem("Record from microphone"):
|
184 |
+
microphone = gr.Audio(
|
185 |
source="microphone", # Choose between "microphone", "upload"
|
186 |
type="filepath",
|
187 |
optional=False,
|
188 |
label="Record from microphone",
|
189 |
)
|
|
|
|
|
|
|
190 |
|
191 |
record_button = gr.Button("Submit for recognition")
|
192 |
+
recorded_output = gr.HTML(label="Recognized speech from recordings")
|
193 |
|
194 |
upload_button.click(
|
195 |
process,
|