app

app.py

@@ -206,14 +206,16 @@ def show_generation_result(model, text, audio_path, motion_length, result_path):
     if audio_path is not None:
         add_audio(result_path, [audio_path])
         
-def generate(prompt, audio_path, length):
+def generate(prompt, audio_check, audio_path, length):
     if not os.path.exists("outputs"):
         os.mkdir("outputs")
     result_path = "outputs/" + str(int(time.time())) + ".mp4"
     print(audio_path)
-    if not os.path.exists(audio_path):
+    if not audio_check:
         audio_path = None
-    if audio_path.endswith("placeholder.wav"):
+    if audio_path is not None and not os.path.exists(audio_path):
+        audio_path = None
+    if audio_path is not None and audio_path.endswith("placeholder.wav"):
         audio_path = None
     if len(prompt) == 0: 
         prompt = None
@@ -240,19 +242,20 @@ input_text = gr.Textbox(
     label="Text prompt:"
 )
 
+audio_check = gr.Checkbox(
+    label="Enable audio? "
+)
+
 demo = gr.Interface(
     fn=generate,
-    inputs=[input_text, input_audio, gr.Slider(20, 200, value=60, label="Motion length (fps 20):")],
+    inputs=[input_text, audio_check, input_audio, gr.Slider(20, 200, value=60, label="Motion length (fps 20):")],
     outputs=gr.Video(label="Video:"),
     examples=[
-        # ["A person walks in a circle.", load_file_list["audio_placeholder"], 120],
-        # ["A person jumps forward.", load_file_list["audio_placeholder"], 100],
-        # ["A person is stretching arms.", load_file_list["audio_placeholder"], 80],
-        ["A person walks in a circle.", "None", 120],
-        ["A person jumps forward.", "None", 100],
-        ["A person is stretching arms.", "None", 80],
-        ["", load_file_list["audio_surprise"], 200],
-        ["", load_file_list["audio_angry"], 200],
+        ["A person walks in a circle.", False, load_file_list["audio_placeholder"], 120],
+        ["A person jumps forward.", False, load_file_list["audio_placeholder"], 100],
+        ["A person is stretching arms.", False, load_file_list["audio_placeholder"], 80],
+        ["", True, load_file_list["audio_surprise"], 200],
+        ["", True, load_file_list["audio_angry"], 200],
     ],
     title="LMM: Large Motion Model for Unified Multi-Modal Motion Generation",
     description="\nThis is an interactive demo for LMM. For more information, feel free to visit our project page(https://github.com/mingyuan-zhang/LMM).")