Fitting Mesh With MVs error.

[jakechai]

RuntimeError: Inference tensors cannot be saved for backward. To work around you can make a clone to get a normal tensor and use it in autograd.

got prompt [load_obj] use texture from: E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\output\warrior\Unique3D\Mesh-InstantM\warrior_albedo.png [load_obj] load texture: (2048, 2048, 3) [Mesh loading] v: torch.Size([54340, 3]), f: torch.Size([108664, 3]) 0%| | 0/2048 [00:00<?, ?it/s] !!! Exception during processing!!! Inference tensors cannot be saved for backward. To work around you can make a clone to get a normal tensor and use it in autograd. Traceback (most recent call last): File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\custom_nodes\ComfyUI-3D-Pack\nodes.py", line 1283, in fitting_mesh mesh_fitter.training() File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\custom_nodes\ComfyUI-3D-Pack\MVs_Algorithms\DiffRastMesh\diff_mesh.py", line 104, in training out = self.cam_controller.render_at_pose(self.all_ref_cam_poses[i]) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\custom_nodes\ComfyUI-3D-Pack\shared_utils\camera_utils.py", line 235, in render_at_pose return self.get_render_result(render_pose, bg_color, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\custom_nodes\ComfyUI-3D-Pack\MVs_Algorithms\DiffRastMesh\diff_mesh.py", line 22, in get_render_result return self.renderer.render(*ref_cam, self.cam.H, self.cam.W, ssaa=1, bg_color=bg_color, **kwargs) #ssaa = min(2.0, max(0.125, 2 * np.random.random())) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\ComfyUI\custom_nodes\ComfyUI-3D-Pack\MVs_Algorithms\DiffRastMesh\diff_mesh_renderer.py", line 139, in render albedo = dr.antialias(albedo, rast, v_clip, self.mesh.f).squeeze(0).contiguous() # [H, W, 3] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\python_embeded\Lib\site-packages\nvdiffrast\torch\ops.py", line 702, in antialias return _antialias_func.apply(color, rast, pos, tri, topology_hash, pos_gradient_boost) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "E:\Program Files\ComfyUI_windows_portable_3D\python_embeded\Lib\site-packages\torch\autograd\function.py", line 598, in apply return super().apply(*args, **kwargs) # type: ignore[misc] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: Inference tensors cannot be saved for backward. To work around you can make a clone to get a normal tensor and use it in autograd.