init project

app.py

@@ -93,6 +93,7 @@ def _convert_scene_output_to_glb(outdir, imgs, pts3d, mask, focals, cams2world,
     scene.export(file_obj=outfile)
     return outfile
 
+
 def get_3D_model_from_scene(outdir, scene, min_conf_thr=3, as_pointcloud=False, mask_sky=False,
                             clean_depth=False, transparent_cams=False, cam_size=0.05):
     """
@@ -447,7 +448,7 @@ def get_cog_feats(images, sam2, siglip, siglip_processor, yolov8, mobilesamv2):
     return cog_seg_maps, rev_cog_seg_maps, multi_view_clip_feats
 
 
-@spaces.GPU(duration=30)
+@spaces.GPU(duration=120)
 def get_reconstructed_scene(outdir, filelist, schedule='linear', niter=300, min_conf_thr=3.0,
                             as_pointcloud=True, mask_sky=False, clean_depth=True, transparent_cams=True, cam_size=0.05,
                             scenegraph_type='complete', winsize=1, refid=0):
@@ -540,12 +541,10 @@ def get_reconstructed_scene(outdir, filelist, schedule='linear', niter=300, min_
     torch.cuda.empty_cache()
 
 
-
     return outfile 
 
-
-# def get_3D_object_from_scene(outdir, text, threshold, scene, min_conf_thr, as_pointcloud, 
-#                  mask_sky, clean_depth, transparent_cams, cam_size):
+# def get_3D_object_from_scene(outdir, text, threshold, scene, min_conf_thr=3.0, as_pointcloud=True, 
+#                  mask_sky=False, clean_depth=True, transparent_cams=True, cam_size=0.05):
     
 #     device = 'cpu'
 #     siglip_tokenizer = AutoTokenizer.from_pretrained("google/siglip-large-patch16-256")
@@ -559,9 +558,22 @@ def get_reconstructed_scene(outdir, filelist, schedule='linear', niter=300, min_
 #         text_feats = text_feats / text_feats.norm(dim=-1, keepdim=True)
 #     scene.render_image(text_feats, threshold)
 #     scene.ori_imgs = scene.rendered_imgs
-#     outfile = get_3D_model_from_scene(outdir, scene, min_conf_thr, as_pointcloud, mask_sky,
-#                                       clean_depth, transparent_cams, cam_size)
-#     return outfile
+
+
+#     rgbimg = scene.ori_imgs
+#     focals = scene.get_focals().cpu()
+#     cams2world = scene.get_im_poses().cpu()
+#     # 3D pointcloud from depthmap, poses and intrinsics
+#     pts3d = to_numpy(scene.get_pts3d())
+#     scene.min_conf_thr = float(scene.conf_trf(torch.tensor(min_conf_thr)))
+#     msk = to_numpy(scene.get_masks())
+#     return _convert_scene_output_to_glb(outdir, rgbimg, pts3d, msk, focals, cams2world, as_pointcloud=as_pointcloud,
+#                                         transparent_cams=transparent_cams, cam_size=cam_size)
+
+
+
+
+
 
 tmpdirname = tempfile.mkdtemp(suffix='pe3r_gradio_demo')
 
@@ -593,7 +605,6 @@ with gradio.Blocks(css=""".gradio-container {margin: 0 !important; min-width: 10
                         outputs=[outmodel]) # , outgallery, scene, 
         
         # find_btn.click(fn=get_3D_object_from_scene_fun,
-        #                     inputs=[text_input, threshold, scene, min_conf_thr, as_pointcloud, mask_sky,
-        #                             clean_depth, transparent_cams, cam_size],
+        #                     inputs=[text_input, threshold, scene],
         #                 outputs=outmodel)
 demo.launch(show_error=True, share=None, server_name=None, server_port=None)