Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -1,45 +1,108 @@
|
|
|
|
1 |
import argparse
|
2 |
-
|
|
|
3 |
from pipeline_ace_step import ACEStepPipeline
|
4 |
from data_sampler import DataSampler
|
5 |
-
import os
|
6 |
-
|
7 |
-
|
8 |
-
parser = argparse.ArgumentParser()
|
9 |
-
parser.add_argument("--checkpoint_path", type=str, default=None)
|
10 |
-
parser.add_argument("--server_name", type=str, default="0.0.0.0")
|
11 |
-
parser.add_argument("--port", type=int, default=7860)
|
12 |
-
parser.add_argument("--device_id", type=int, default=0)
|
13 |
-
parser.add_argument("--share", action='store_true', default=False)
|
14 |
-
parser.add_argument("--bf16", action='store_true', default=True)
|
15 |
-
parser.add_argument("--torch_compile", type=bool, default=False)
|
16 |
-
|
17 |
-
args = parser.parse_args()
|
18 |
-
os.environ["CUDA_VISIBLE_DEVICES"] = str(args.device_id)
|
19 |
-
|
20 |
|
21 |
-
|
|
|
|
|
|
|
|
|
|
|
22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
-
|
25 |
-
|
|
|
|
|
|
|
|
|
26 |
model_demo = ACEStepPipeline(
|
27 |
-
checkpoint_dir=args
|
28 |
-
dtype="bfloat16" if args
|
29 |
persistent_storage_path=persistent_storage_path,
|
30 |
-
torch_compile=args
|
31 |
)
|
32 |
data_sampler = DataSampler()
|
|
|
33 |
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
|
|
40 |
|
41 |
-
|
42 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
|
44 |
if __name__ == "__main__":
|
45 |
-
main(
|
|
|
1 |
+
# app.py
|
2 |
import argparse
|
3 |
+
import streamlit as st
|
4 |
+
import os
|
5 |
from pipeline_ace_step import ACEStepPipeline
|
6 |
from data_sampler import DataSampler
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
8 |
+
# Streamlit 설정
|
9 |
+
st.set_page_config(
|
10 |
+
page_title="ACE Step Music Generator",
|
11 |
+
page_icon="🎵",
|
12 |
+
layout="wide"
|
13 |
+
)
|
14 |
|
15 |
+
def get_args():
|
16 |
+
"""환경변수 또는 기본값으로 설정"""
|
17 |
+
return {
|
18 |
+
'checkpoint_path': os.environ.get('CHECKPOINT_PATH'),
|
19 |
+
'device_id': int(os.environ.get('DEVICE_ID', '0')),
|
20 |
+
'bf16': os.environ.get('BF16', 'True').lower() == 'true',
|
21 |
+
'torch_compile': os.environ.get('TORCH_COMPILE', 'False').lower() == 'true'
|
22 |
+
}
|
23 |
|
24 |
+
@st.cache_resource
|
25 |
+
def load_model(args):
|
26 |
+
"""모델 로딩 (캐시됨)"""
|
27 |
+
os.environ["CUDA_VISIBLE_DEVICES"] = str(args['device_id'])
|
28 |
+
persistent_storage_path = "/data"
|
29 |
+
|
30 |
model_demo = ACEStepPipeline(
|
31 |
+
checkpoint_dir=args['checkpoint_path'],
|
32 |
+
dtype="bfloat16" if args['bf16'] else "float32",
|
33 |
persistent_storage_path=persistent_storage_path,
|
34 |
+
torch_compile=args['torch_compile']
|
35 |
)
|
36 |
data_sampler = DataSampler()
|
37 |
+
return model_demo, data_sampler
|
38 |
|
39 |
+
def main():
|
40 |
+
st.title("🎵 ACE Step Music Generator")
|
41 |
+
|
42 |
+
args = get_args()
|
43 |
+
|
44 |
+
try:
|
45 |
+
model_demo, data_sampler = load_model(args)
|
46 |
|
47 |
+
# UI 구성
|
48 |
+
col1, col2 = st.columns([2, 1])
|
49 |
+
|
50 |
+
with col1:
|
51 |
+
st.header("Generate Music")
|
52 |
+
|
53 |
+
# 텍스트 입력
|
54 |
+
prompt = st.text_area(
|
55 |
+
"Enter your music description:",
|
56 |
+
placeholder="Enter a description of the music you want to generate...",
|
57 |
+
height=100
|
58 |
+
)
|
59 |
+
|
60 |
+
# 생성 버튼
|
61 |
+
if st.button("Generate Music", type="primary"):
|
62 |
+
if prompt:
|
63 |
+
with st.spinner("Generating music..."):
|
64 |
+
try:
|
65 |
+
result = model_demo(prompt)
|
66 |
+
st.success("Music generated successfully!")
|
67 |
+
|
68 |
+
# 결과 표시 (result 형태에 따라 조정 필요)
|
69 |
+
if hasattr(result, 'audio'):
|
70 |
+
st.audio(result.audio)
|
71 |
+
else:
|
72 |
+
st.write(result)
|
73 |
+
|
74 |
+
except Exception as e:
|
75 |
+
st.error(f"Error generating music: {str(e)}")
|
76 |
+
else:
|
77 |
+
st.warning("Please enter a description first.")
|
78 |
+
|
79 |
+
with col2:
|
80 |
+
st.header("Sample Data")
|
81 |
+
|
82 |
+
if st.button("Load Sample"):
|
83 |
+
try:
|
84 |
+
sample_data = data_sampler.sample()
|
85 |
+
st.json(sample_data)
|
86 |
+
except Exception as e:
|
87 |
+
st.error(f"Error loading sample: {str(e)}")
|
88 |
+
|
89 |
+
# 파일 업로드
|
90 |
+
uploaded_file = st.file_uploader(
|
91 |
+
"Upload JSON data",
|
92 |
+
type=['json']
|
93 |
+
)
|
94 |
+
|
95 |
+
if uploaded_file:
|
96 |
+
try:
|
97 |
+
data = data_sampler.load_json(uploaded_file)
|
98 |
+
st.json(data)
|
99 |
+
except Exception as e:
|
100 |
+
st.error(f"Error loading file: {str(e)}")
|
101 |
+
|
102 |
+
except Exception as e:
|
103 |
+
st.error(f"Error loading model: {str(e)}")
|
104 |
+
import traceback
|
105 |
+
st.code(traceback.format_exc())
|
106 |
|
107 |
if __name__ == "__main__":
|
108 |
+
main()
|