Spaces:
Running
on
Zero
Running
on
Zero
Commit
•
7bd176a
1
Parent(s):
4a0236d
Modify
Browse files
app.py
CHANGED
@@ -65,6 +65,7 @@ frame_rate = model.audio_encoder.config.frame_rate
|
|
65 |
|
66 |
|
67 |
def generate_response(audio):
|
|
|
68 |
gr.Info("Transcribing Audio", duration=5)
|
69 |
question = client.automatic_speech_recognition(audio).text
|
70 |
|
@@ -73,11 +74,11 @@ def generate_response(audio):
|
|
73 |
"is to answer with a cryptic addage or proverb such as "
|
74 |
"'curiosity killed the cat' or 'The early bird gets the worm'."
|
75 |
"Keep your answers short.")},
|
76 |
-
{"role": "user", "content": f"
|
77 |
|
78 |
response = client.chat_completion(messages, max_tokens=64, seed=random.randint(1, 5000))
|
79 |
response = response.choices[0].message.content
|
80 |
-
|
81 |
|
82 |
@spaces.GPU
|
83 |
def generate_base(answer):
|
@@ -106,7 +107,7 @@ def generate_base(answer):
|
|
106 |
|
107 |
for new_audio in streamer:
|
108 |
print(f"Sample of length: {round(new_audio.shape[0] / sampling_rate, 2)} seconds")
|
109 |
-
yield
|
110 |
|
111 |
css=""".my-group {max-width: 600px !important; max-height: 600 !important;}
|
112 |
.my-column {display: flex !important; justify-content: center !important; align-items: center !important};"""
|
@@ -128,6 +129,6 @@ with gr.Blocks() as block:
|
|
128 |
with gr.Row():
|
129 |
audio_in = gr.Audio(label="Speak you question", sources="microphone", type="filepath")
|
130 |
|
131 |
-
audio_in.stop_recording(generate_response, audio_in,
|
132 |
|
133 |
block.launch()
|
|
|
65 |
|
66 |
|
67 |
def generate_response(audio):
|
68 |
+
yield None, None
|
69 |
gr.Info("Transcribing Audio", duration=5)
|
70 |
question = client.automatic_speech_recognition(audio).text
|
71 |
|
|
|
74 |
"is to answer with a cryptic addage or proverb such as "
|
75 |
"'curiosity killed the cat' or 'The early bird gets the worm'."
|
76 |
"Keep your answers short.")},
|
77 |
+
{"role": "user", "content": f"Please tell me what to do about {question}"}]
|
78 |
|
79 |
response = client.chat_completion(messages, max_tokens=64, seed=random.randint(1, 5000))
|
80 |
response = response.choices[0].message.content
|
81 |
+
yield response, None
|
82 |
|
83 |
@spaces.GPU
|
84 |
def generate_base(answer):
|
|
|
107 |
|
108 |
for new_audio in streamer:
|
109 |
print(f"Sample of length: {round(new_audio.shape[0] / sampling_rate, 2)} seconds")
|
110 |
+
yield answer, numpy_to_mp3(new_audio, sampling_rate=sampling_rate)
|
111 |
|
112 |
css=""".my-group {max-width: 600px !important; max-height: 600 !important;}
|
113 |
.my-column {display: flex !important; justify-content: center !important; align-items: center !important};"""
|
|
|
129 |
with gr.Row():
|
130 |
audio_in = gr.Audio(label="Speak you question", sources="microphone", type="filepath")
|
131 |
|
132 |
+
audio_in.stop_recording(generate_response, audio_in, state).then(fn=generate_base, inputs=state, outputs=[answer, audio_out])
|
133 |
|
134 |
block.launch()
|