Spaces:
Runtime error
Runtime error
Added app
Browse files
app.py
ADDED
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import nemo.collections.asr as nemo_asr
|
2 |
+
import gradio as gr
|
3 |
+
|
4 |
+
|
5 |
+
|
6 |
+
|
7 |
+
asr_model = nemo_asr.models.ASRModel.from_pretrained("theodotus/stt_ua_fastconformer_hybrid_large_pc", map_location="cpu")
|
8 |
+
|
9 |
+
|
10 |
+
|
11 |
+
|
12 |
+
def process_file(in_filename: str,):
|
13 |
+
if in_filename is None or in_filename == "":
|
14 |
+
return "Error: No file"
|
15 |
+
|
16 |
+
transcript = asr_model.transcribe([in_filename])[0][0]
|
17 |
+
|
18 |
+
|
19 |
+
return transcript
|
20 |
+
|
21 |
+
|
22 |
+
|
23 |
+
|
24 |
+
demo = gr.Blocks()
|
25 |
+
|
26 |
+
with demo:
|
27 |
+
with gr.Tabs():
|
28 |
+
with gr.TabItem("Upload from disk"):
|
29 |
+
uploaded_file = gr.Audio(
|
30 |
+
source="upload", # Choose between "microphone", "upload"
|
31 |
+
type="filepath",
|
32 |
+
optional=False,
|
33 |
+
label="Upload from disk",
|
34 |
+
)
|
35 |
+
upload_button = gr.Button("Submit for recognition")
|
36 |
+
uploaded_output = gr.Textbox(label="Recognized speech from uploaded file")
|
37 |
+
|
38 |
+
with gr.TabItem("Record from microphone"):
|
39 |
+
microphone = gr.Audio(
|
40 |
+
source="microphone", # Choose between "microphone", "upload"
|
41 |
+
type="filepath",
|
42 |
+
optional=False,
|
43 |
+
label="Record from microphone",
|
44 |
+
)
|
45 |
+
|
46 |
+
record_button = gr.Button("Submit for recognition")
|
47 |
+
recorded_output = gr.Textbox(label="Recognized speech from recordings")
|
48 |
+
|
49 |
+
upload_button.click(
|
50 |
+
process_file,
|
51 |
+
inputs=[
|
52 |
+
uploaded_file,
|
53 |
+
],
|
54 |
+
outputs=[uploaded_output],
|
55 |
+
)
|
56 |
+
|
57 |
+
record_button.click(
|
58 |
+
process_file,
|
59 |
+
inputs=[
|
60 |
+
microphone,
|
61 |
+
],
|
62 |
+
outputs=[recorded_output],
|
63 |
+
)
|
64 |
+
|
65 |
+
|
66 |
+
if __name__ == "__main__":
|
67 |
+
demo.launch()
|