Spaces:
Running
on
L40S
Running
on
L40S
Update app.py
Browse files
app.py
CHANGED
@@ -7,6 +7,31 @@ import json
|
|
7 |
import tempfile
|
8 |
from huggingface_hub import snapshot_download
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
num_gpus = torch.cuda.device_count()
|
11 |
print(f"GPU AVAILABLE: {num_gpus}")
|
12 |
|
@@ -114,6 +139,11 @@ def create_temp_input_json(prompt: str, cond_image_path: str, cond_audio_path: s
|
|
114 |
|
115 |
|
116 |
def infer(prompt, cond_image_path, cond_audio_path):
|
|
|
|
|
|
|
|
|
|
|
117 |
# Prepare input JSON
|
118 |
input_json_path = create_temp_input_json(prompt, cond_image_path, cond_audio_path)
|
119 |
|
@@ -140,24 +170,29 @@ def infer(prompt, cond_image_path, cond_audio_path):
|
|
140 |
else:
|
141 |
cmd = ["python3", "generate_multitalk.py"] + common_args
|
142 |
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
156 |
-
|
157 |
-
|
158 |
-
|
159 |
-
|
160 |
-
|
|
|
|
|
|
|
|
|
|
|
161 |
|
162 |
|
163 |
with gr.Blocks(title="MultiTalk Inference") as demo:
|
|
|
7 |
import tempfile
|
8 |
from huggingface_hub import snapshot_download
|
9 |
|
10 |
+
import soundfile as sf
|
11 |
+
import tempfile
|
12 |
+
from datetime import datetime
|
13 |
+
|
14 |
+
is_shared_ui = True if "fffiloni/Meigen-MultiTalk" in os.environ['SPACE_ID'] else False
|
15 |
+
|
16 |
+
def trim_audio_to_5s_temp(audio_path, sample_rate=16000):
|
17 |
+
max_duration_sec = 5
|
18 |
+
audio, sr = sf.read(audio_path)
|
19 |
+
|
20 |
+
if sr != sample_rate:
|
21 |
+
raise ValueError(f"Expected sample rate {sample_rate}, but got {sr}")
|
22 |
+
|
23 |
+
max_samples = max_duration_sec * sample_rate
|
24 |
+
if len(audio) > max_samples:
|
25 |
+
audio = audio[:max_samples]
|
26 |
+
|
27 |
+
timestamp = datetime.now().strftime("%Y%m%d%H%M%S%f")
|
28 |
+
base_name = os.path.splitext(os.path.basename(audio_path))[0]
|
29 |
+
temp_filename = f"{base_name}_trimmed_{timestamp}.wav"
|
30 |
+
temp_path = os.path.join(tempfile.gettempdir(), temp_filename)
|
31 |
+
|
32 |
+
sf.write(temp_path, audio, samplerate=sample_rate)
|
33 |
+
return temp_path
|
34 |
+
|
35 |
num_gpus = torch.cuda.device_count()
|
36 |
print(f"GPU AVAILABLE: {num_gpus}")
|
37 |
|
|
|
139 |
|
140 |
|
141 |
def infer(prompt, cond_image_path, cond_audio_path):
|
142 |
+
|
143 |
+
if is_shared_ui:
|
144 |
+
trimmed_audio_path = trim_audio_to_5s_temp(cond_audio_path)
|
145 |
+
cond_audio_path = trimmed_audio_path
|
146 |
+
|
147 |
# Prepare input JSON
|
148 |
input_json_path = create_temp_input_json(prompt, cond_image_path, cond_audio_path)
|
149 |
|
|
|
170 |
else:
|
171 |
cmd = ["python3", "generate_multitalk.py"] + common_args
|
172 |
|
173 |
+
try:
|
174 |
+
# Log to file and stream
|
175 |
+
with open("inference.log", "w") as log_file:
|
176 |
+
process = subprocess.Popen(
|
177 |
+
cmd,
|
178 |
+
stdout=subprocess.PIPE,
|
179 |
+
stderr=subprocess.STDOUT,
|
180 |
+
text=True,
|
181 |
+
bufsize=1
|
182 |
+
)
|
183 |
+
for line in process.stdout:
|
184 |
+
print(line, end="")
|
185 |
+
log_file.write(line)
|
186 |
+
process.wait()
|
187 |
+
|
188 |
+
if process.returncode != 0:
|
189 |
+
raise RuntimeError("Inference failed. Check inference.log for details.")
|
190 |
+
|
191 |
+
return "multi_long_multigpu_exp.mp4"
|
192 |
+
|
193 |
+
finally:
|
194 |
+
if os.path.exists(trimmed_audio_path):
|
195 |
+
os.remove(trimmed_audio_path)
|
196 |
|
197 |
|
198 |
with gr.Blocks(title="MultiTalk Inference") as demo:
|