config file

app.py

@@ -13,38 +13,42 @@ import torch
 from diffusers import MochiPipeline
 from diffusers.utils import export_to_video
 import gradio as gr
+import config as cfg
 
-# Caricare il modello pre-addestrato
-pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview", variant="bf16", torch_dtype=torch.bfloat16)
+# Load the pre-trained model
+pipe = MochiPipeline.from_pretrained(cfg.MODEL_PRE_TRAINED_ID, variant="bf16", torch_dtype=torch.bfloat16)
 
-# Abilitare le ottimizzazioni per il risparmio di memoria
+# Enable memory-saving optimizations
 pipe.enable_model_cpu_offload()
 pipe.enable_vae_tiling()
 
+@spaces.GPU(duration=600)
+def generate_video(prompt, num_frames=84, fps=30):
+    # Generate video frames
+    print("Generating video frames...")
+    frames = pipe(prompt, num_frames=num_frames).frames[0]
 
-@spaces.GPU(duration=200)
-def generate_video(prompt):
-    # Generare i frame del video
-    print("Generare i frame del video...")
-    #frames = pipe(prompt, num_frames=84).frames[0]
-    frames = pipe(prompt, num_frames=24).frames[0]
-    
-    # Esportare i frame come video
+    # Export frames as video
     video_path = "mochi.mp4"
-    export_to_video(frames, video_path, fps=30)
-
+    export_to_video(frames, video_path, fps=fps)
     return video_path
 
-
-# Creare l'interfaccia Gradio
+# Create the Gradio interface
 interface = gr.Interface(
     fn=generate_video,
-    inputs="text",
-    outputs="video",
-    title="Mochi Video Generator",
-    description="Genera un video basato su un prompt di testo utilizzando MochiPipeline."
+    inputs=[
+        gr.inputs.Textbox(lines=2, placeholder="Enter your text prompt here... 💡"),
+        gr.inputs.Slider(minimum=1, maximum=240, default=84, label="Number of frames 🎞️"),
+        gr.inputs.Slider(minimum=1, maximum=60, default=30, label="FPS (Frames per second) ⏱️")
+    ],
+    outputs=gr.outputs.Video(),
+    title=cfg.TITLE,
+    description=cfg.DESCRIPTION,
+    examples=cfg.EXAMPLES,
+    article=cfg.BUY_ME_A_COFFEE
+
 )
 
-# Avviare l'applicazione
+# Launch the application
 if __name__ == "__main__":
     interface.launch()

config.py

@@ -0,0 +1,39 @@
+import os
+import numpy as np
+
+TITLE = "Mochi 1 - Video Generator 🌟"
+
+DESCRIPTION = """  
+✨ Generate a video based on a text prompt using the MochiPipeline. 🎥                                            
+Enter a text prompt and customize the number of frames and the frame rate (fps) for the generated video. 🛠️     
+"""
+
+BUY_ME_A_COFFE= """  
+<a href="https://buymeacoffee.com/thesab" target="_blank">  
+    <button style="background-color: #FFDD00; border: none; color: black; padding: 10px 20px; text-align: center; text-decoration: none; display: inline-block; font-size: 16px; margin: 4px 2px; cursor: pointer; border-radius: 10px;">  
+        ☕ Buy Me a Coffee  
+    </button>  
+</a>   
+"""
+
+
+MODEL_PRE_TRAINED_ID = "genmo/mochi-1-preview"
+
+MAX_SEED = np.iinfo(np.int32).max
+MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "1024"))
+DEFAULT_SEED = 42
+DEFAULT_RANDOMIZE_SEED = False
+DEFAULT_WIDTH = 1024
+DEFAULT_HEIGHT = 1024
+DEFAULT_GUIDANCE_SCALE = 4.5
+DEFAULT_NUM_INFERENCE_STEPS = 40
+DEFAULT_LORA_SCALE = 4.5
+
+EXAMPLES = [
+     ["A beautiful sunrise over the mountains 🌄", 100, 24],
+    ["A futuristic city with flying cars 🚗✨", 120, 30],
+    ["A serene beach with waves crashing 🌊", 80, 25],
+    ["A bustling market in a medieval town 🏰", 90, 20],
+    ["A spaceship traveling through space 🚀", 110, 30]
+]
+