Enable ZeroGPU for SAM 3D Objects

README.md

@@ -7,6 +7,7 @@ sdk: gradio
 sdk_version: 5.49.1
 app_file: app.py
 pinned: false
+suggested_hardware: zero-a10g
 license: other
 short_description: Object-centric 3D reconstruction with Meta SAM 3D Objects
 ---
@@ -28,6 +29,7 @@ Output:
 Operational notes:
 
 - The official setup requires a Linux NVIDIA GPU with at least 32 GB VRAM.
+- This Space is configured to try Hugging Face ZeroGPU first.
 - The model checkpoints are gated. The Space needs an `HF_TOKEN` secret that has access to `facebook/sam-3d-objects`.
 - If the Space starts on CPU or a small GPU, the app will show diagnostics instead of failing silently.
 

app.py

@@ -8,6 +8,7 @@ from pathlib import Path
 import gradio as gr
 import numpy as np
 from PIL import Image
+import spaces
 
 
 MODEL_REPO = "facebook/sam-3d-objects"
@@ -108,6 +109,7 @@ def prepare_mask(mask_image):
     return mask
 
 
+@spaces.GPU(duration=300)
 def reconstruct(image, mask_image, seed):
     if image is None:
         raise gr.Error("Upload an input image.")