李化宣

Results 1 comments of 李化宣

尝试修改了一下,当我仅输入图像不输入视频位姿的时候,渲染的结果出不来,人体的3D高斯生成不了,query和latent_features正常 gs_model_list, query_points, smplx_params = self.renderer.forward_gs( gs_hidden_features=latent_points, query_points=query_points, smplx_data=smplx_params, additional_features={"image_feats": image_feats, "image": image[:, 0]}, )