WorldFoundry
English中文

3D & 4D · 3d-reconstruction · gaussian-splatting

Splatt3R

brandonsmart

Runtime code is vendored in-tree under worldfoundry/base_models/three_dimensions/general_3d/splatt3r/splatt3r_runtime; checkpoints remain external assets.

Runner verifiedUnified environmentsplatt3r

Build your run

Choose a recorded variant, then copy the exact setup or inspection command.

worldfoundry.pipeline
Task3d-reconstruction
Environmentworldfoundry-unified-cu128
DeviceCUDA 11.3
worldfoundry-eval evaluate \
  --mode model \
  --model-id splatt3r \
  --model-runner worldfoundry:pipeline \
  --model-manifest-dir worldfoundry/data/models/catalog \
  --requests-path tmp/requests.jsonl \
  --output-dir tmp/model_eval/splatt3r \
  --metric artifact_count \
  --json
01

Compatibility & versions

Manifest-backed recipe

Integration
verified
Runner evidence
verified
Environment
worldfoundry-unified-cu128
Python
3.10
CUDA
CUDA 11.3
PyTorch
torch
Source revision
—
Checkpoint revision
—
Runtime profile
splatt3r
Pipeline binding
splatt3r
Runner
worldfoundry.pipeline
Pipeline target
worldfoundry.pipelines.component_pipelines:Splatt3RPipeline
Backend stage
in_tree_runtime
Runtime status
checkpoint_verified
Driver status
compatible_with_downgraded_torch
Environment kind
Unified environment
02

Install environment

The environment resolver reads the recorded profile and chooses the unified or dedicated environment shown here.

bash scripts/setup/model_env_install.sh --model splatt3r
Package constraints 15
  • torch
  • torchvision
  • numpy
  • einops
  • omegaconf
  • opencv-python
  • pillow
  • scipy
  • timm
  • trimesh
  • plyfile
  • gradio
  • socksio
  • huggingface_hub
  • safetensors
Conda packages 3
  • python
  • pip
  • ffmpeg
03

Checkpoints & assets

Run the local check before allocating compute. Gated, private, and license fields below come directly from the checkpoint manifest.

worldfoundry-eval zoo model-download --model-id splatt3r --check-local --json
brandonsmart/splatt3r_v1.0
Revision
—
License
—
Gated
—
Private
—
04

Launch & outputs

The generated command uses the shared evaluation boundary and writes durable result manifests and artifacts.

worldfoundry-eval evaluate \
  --mode model \
  --model-id splatt3r \
  --model-runner worldfoundry:pipeline \
  --model-manifest-dir worldfoundry/data/models/catalog \
  --requests-path tmp/requests.jsonl \
  --output-dir tmp/model_eval/splatt3r \
  --metric artifact_count \
  --json

Input contract

FieldRecorded contract
actionsview_sequence
imageRequired
promptOptional

Artifact contract

Artifact kindFilename / path
generated_3d_assetsplatt3r.ply
05

Evidence & sources

Catalog integration, native-demo parity, and runner parity are independent records.

Integration
verified
Runner evidence
verified
Native demo evidence
Not recorded
Validation imports
torch, numpy, trimesh, gradio
Recipe provenance
worldfoundry/data/models/catalog/three_d_four_d/splatt3r.yamlProjecthttps://splatt3r.active.vision/Paperhttps://arxiv.org/abs/2408.13912GitHubhttps://github.com/btsmart/splatt3rHugging Facehttps://huggingface.co/brandonsmart/splatt3r_v1.0