World models · world · camera_control · world_model

LingBot-World

LingBot

LingBot-World wrapper now only resolves local HFD or explicit checkpoint directories and no longer downloads at runtime.

Runner verifiedDedicated environmentlingbot-world

Build your run

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

worldfoundry.pipeline

Variant

Taskworld

Environmentlingbot-world

DeviceCUDA 12.1

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

Compatibility & versions

Manifest-backed recipe

Integration: verified
Runner evidence: verified
Environment: lingbot-world
Python: 3.10
CUDA: CUDA 12.1
PyTorch: torch==2.5.1
Source revision: —
Checkpoint revision: 6fc824ffc338d64c97c77e2eb8c0f4cfc24d82bd

Runtime profile: lingbot-world
Pipeline binding: lingbot-world
Runner: worldfoundry.pipeline
Pipeline target: worldfoundry.pipelines.lingbot_world.pipeline_lingbot_world:LingBotPipeline
Backend stage: in_tree_runtime
Runtime status: base_cam_contract
Driver status: compatible
Environment kind: Dedicated environment

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 lingbot-world

Package constraints 18

torch==2.5.1
torchvision==0.20.1
flash-attn==2.8.3
diffusers>=0.31.0
transformers>=4.49.0,<=4.51.3
tokenizers>=0.20.3
accelerate>=1.1.1
opencv-python>=4.9.0.80
imageio[ffmpeg]
imageio-ffmpeg
einops
easydict
ftfy
numpy>=1.23.5,<2
scipy
tqdm
pillow
pyyaml

Conda packages 3

python=3.10
pip
ffmpeg

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 lingbot-world --check-local --json

robbyant/lingbot-world-base-cam

Revision: 6fc824ffc338d64c97c77e2eb8c0f4cfc24d82bd
License: —
Gated: —
Private: —

Launch & outputs

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

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

Input contract

Field	Recorded contract
`prompt`	Required
`image`	Required
`video`	Optional
`actions`	camera_path, navigation_action

Artifact contract

Artifact kind	Filename / path
`generated_world`	`lingbot-world.mp4`

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, flash_attn, cv2, imageio

Recipe provenance

worldfoundry/data/models/catalog/world_models/lingbot-world.yamlGitHubhttps://github.com/robbyant/lingbot-worldHugging Facehttps://huggingface.co/robbyant/lingbot-world-base-cam

Build your run

Input contract

Artifact contract

Related recipes