Update app.py
Browse files
app.py
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
from audiocraft.models import MusicGen
|
2 |
import streamlit as st
|
3 |
-
import os
|
4 |
import torch
|
5 |
import torchaudio
|
6 |
import numpy as np
|
@@ -8,13 +8,13 @@ import base64
|
|
8 |
|
9 |
@st.cache_resource
|
10 |
def load_model():
|
11 |
-
model=MusicGen.get_pretrained("facebook/musicgen-small")
|
12 |
return model
|
13 |
|
14 |
-
def generate_music_tensors(description,duration:int):
|
15 |
-
print("Description:",description)
|
16 |
-
print("Duration:",duration)
|
17 |
-
model=load_model()
|
18 |
|
19 |
model.set_generation_params(
|
20 |
use_sampling=True,
|
@@ -22,31 +22,33 @@ def generate_music_tensors(description,duration:int):
|
|
22 |
duration=duration
|
23 |
)
|
24 |
|
25 |
-
output=model.generate(
|
26 |
descriptions=[description],
|
27 |
progress=True,
|
28 |
return_tokens=True
|
29 |
)
|
30 |
return output[0]
|
31 |
|
32 |
-
def save_audio(samples:torch.
|
33 |
-
sample_rate=32000
|
34 |
-
save_path="audio_output/"
|
|
|
35 |
|
36 |
-
assert samples.dim()==2 or samples.dim()==3
|
37 |
-
samples=samples.detach().cpu()
|
38 |
|
39 |
-
if samples.dim()==2:
|
40 |
-
samples=samples[None
|
41 |
-
for idx,audio in enumerate(samples):
|
42 |
-
audio_path=os.path.join(save_path,f"
|
43 |
-
torchaudio.save(audio_path,audio,sample_rate)
|
|
|
44 |
|
45 |
-
def get_binary_file_downloader_html(bin_file,file_label='File'):
|
46 |
-
with open(bin_file,'rb') as f:
|
47 |
-
data=f.read()
|
48 |
-
bin_str=base64.b64encode(data).decode()
|
49 |
-
href=f'<a href="data:application/octet-stream;base64,{bin_str} download
|
50 |
return href
|
51 |
|
52 |
st.set_page_config(
|
@@ -59,23 +61,24 @@ def main():
|
|
59 |
|
60 |
with st.expander("See Explanation"):
|
61 |
st.write("App is developed by using Meta's Audiocraft Music Gen model. Write your text and we will generate audio")
|
62 |
-
|
63 |
-
|
|
|
64 |
|
65 |
if text_area and time_slider:
|
66 |
st.json(
|
67 |
{
|
68 |
-
"Description":text_area,
|
69 |
-
"Selected duration
|
70 |
}
|
71 |
)
|
72 |
st.subheader("Generated Music")
|
73 |
-
music_tensors=generate_music_tensors(text_area,time_slider)
|
74 |
-
|
75 |
-
audio_file_path
|
76 |
-
audio_file
|
77 |
-
audio_bytes=audio_file.read()
|
78 |
st.audio(audio_bytes)
|
79 |
-
st.markdown(get_binary_file_downloader_html
|
80 |
-
|
|
|
81 |
main()
|
|
|
1 |
from audiocraft.models import MusicGen
|
2 |
import streamlit as st
|
3 |
+
import os
|
4 |
import torch
|
5 |
import torchaudio
|
6 |
import numpy as np
|
|
|
8 |
|
9 |
@st.cache_resource
|
10 |
def load_model():
|
11 |
+
model = MusicGen.get_pretrained("facebook/musicgen-small")
|
12 |
return model
|
13 |
|
14 |
+
def generate_music_tensors(description, duration: int):
|
15 |
+
print("Description:", description)
|
16 |
+
print("Duration:", duration)
|
17 |
+
model = load_model()
|
18 |
|
19 |
model.set_generation_params(
|
20 |
use_sampling=True,
|
|
|
22 |
duration=duration
|
23 |
)
|
24 |
|
25 |
+
output = model.generate(
|
26 |
descriptions=[description],
|
27 |
progress=True,
|
28 |
return_tokens=True
|
29 |
)
|
30 |
return output[0]
|
31 |
|
32 |
+
def save_audio(samples: torch.Tensor):
|
33 |
+
sample_rate = 32000 # corrected to integer
|
34 |
+
save_path = "audio_output/"
|
35 |
+
os.makedirs(save_path, exist_ok=True) # ensure directory exists
|
36 |
|
37 |
+
assert samples.dim() == 2 or samples.dim() == 3
|
38 |
+
samples = samples.detach().cpu()
|
39 |
|
40 |
+
if samples.dim() == 2:
|
41 |
+
samples = samples[None, ...]
|
42 |
+
for idx, audio in enumerate(samples):
|
43 |
+
audio_path = os.path.join(save_path, f"audio_{idx}.wav")
|
44 |
+
torchaudio.save(audio_path, audio, sample_rate)
|
45 |
+
return os.path.join(save_path, "audio_0.wav")
|
46 |
|
47 |
+
def get_binary_file_downloader_html(bin_file, file_label='File'):
|
48 |
+
with open(bin_file, 'rb') as f:
|
49 |
+
data = f.read()
|
50 |
+
bin_str = base64.b64encode(data).decode()
|
51 |
+
href = f'<a href="data:application/octet-stream;base64,{bin_str}" download="{file_label}">Download {file_label} from here</a>'
|
52 |
return href
|
53 |
|
54 |
st.set_page_config(
|
|
|
61 |
|
62 |
with st.expander("See Explanation"):
|
63 |
st.write("App is developed by using Meta's Audiocraft Music Gen model. Write your text and we will generate audio")
|
64 |
+
|
65 |
+
text_area = st.text_area("Enter description")
|
66 |
+
time_slider = st.slider("Select time duration(s)", 2, 5, 20)
|
67 |
|
68 |
if text_area and time_slider:
|
69 |
st.json(
|
70 |
{
|
71 |
+
"Description": text_area,
|
72 |
+
"Selected duration": time_slider
|
73 |
}
|
74 |
)
|
75 |
st.subheader("Generated Music")
|
76 |
+
music_tensors = generate_music_tensors(text_area, time_slider)
|
77 |
+
audio_file_path = save_audio(music_tensors)
|
78 |
+
audio_file = open(audio_file_path, 'rb')
|
79 |
+
audio_bytes = audio_file.read()
|
|
|
80 |
st.audio(audio_bytes)
|
81 |
+
st.markdown(get_binary_file_downloader_html(audio_file_path, 'Audio'), unsafe_allow_html=True)
|
82 |
+
|
83 |
+
if __name__ == "__main__":
|
84 |
main()
|