WorldFoundry
English中文

LoGeR

Method overview: Comparison of different methods across varying sequence lengths and scene scales. Although FastVGGT is able to process a larger number of frames during inference, it fails completely on large-scale scenes,.

LoGeR: 3d-reconstruction and trajectory; Verified.

Edit this pageGitHubMarkdown

LoGeR reconstructs 3D scenes with camera trajectories and point clouds from input views, and is one of the verified 3D routes in this catalog: checkpoint-backed Studio inference produced finite poses, depth, and a 601,570-vertex point cloud, with evidence in inference_evidence.yaml.

Official paper teaser: Overview of a single block of our hybrid memory module. We process the input sequence in consecutive chunks of frames. While each block utilizes frame and bidirectional attention from prior work, we introduce new.
Official paper teaser: Overview of a single block of our hybrid memory module. We process the input sequence in consecutive chunks of frames. While each block utilizes frame and bidirectional attention from prior work, we introduce new.
Method overview: Comparison of different methods across varying sequence lengths and scene scales. Although FastVGGT is able to process a larger number of frames during inference, it fails completely on large-scale scenes,.
Method overview: Comparison of different methods across varying sequence lengths and scene scales. Although FastVGGT is able to process a larger number of frames during inference, it fails completely on large-scale scenes,.

The recorded paper is arXiv 2603.03269.

Project page: LoGeR-project.github.io. Code: Junyi42/LoGeR.

Build your run

Pipeline
LoGeRPipelineworldfoundry.pipelines.pi3.pipeline_loger
TaskInteractive Video interactive-video
Environmentworldfoundry-unified-cu128
DeviceCUDA 12.8
Runnerworldfoundry.pipeline
Loadingfrom_pretrained
Invocationnative_or_standard_invocation
worldfoundry-eval run \
  loger \
  --pipeline.task-profile interactive-video \
  --json

What it is for

Pick loger for feed-forward geometry: views in, reconstruction / trajectory / point cloud out. Use it when you need the catalog 3D route rather than a depth-only prior.

The catalog records a reconstruction-plus-trajectory model exporting binary PLY artifacts; the publishing institution and architecture internals are not recorded in this repository (upstream: Junyi42/LoGeR, apache-2.0).

Catalog tasks include 3d-reconstruction, trajectory, and point-cloud. Documented outputs include point cloud.

Runtime contract

Optional input_path, interactions, task_type, camera_view, and visualize_ops; artifacts result.ply.

CLI task profiles: interactive-video. WorldFoundry binds LoGeRPipeline.

Family status: Verified.

Verified end to end in the shared worldfoundry-unified-cu128 environment — a dependable choice for point-cloud navigation pipelines.

In-tree LoGeR point-cloud navigation pipeline is registered for reconstruction and trajectory rendering.

Running in WorldFoundry

The catalog launch is worldfoundry-eval run loger --pipeline.task-profile interactive-video. The command builder above switches --pipeline.task-profile among recorded tasks and still emits worldfoundry-eval run.

It runs in the unified environment worldfoundry-unified-cu128 (Python 3.11, CUDA 12.8).

Recorded weight repositories include Junyi42/LoGeR. Prepare the environment and check local assets with bash scripts/inference/prepare_model_infer.sh loger. Add --download to fetch missing public checkpoints.

Status: Verified.