World models · image-to-video · camera-control · interactive-world-model
Astra
EvanEternal
In-tree Astra inference route is registered under worldfoundry/synthesis/visual_generation/kling/astra_runtime.
Build your run
Choose a recorded variant, then copy the exact setup or inspection command.
Taskimage-to-video
Environmentworldfoundry-unified-cu128
DeviceCUDA 12.8
worldfoundry-eval evaluate \
--mode model \
--model-id astra \
--model-runner worldfoundry:pipeline \
--model-manifest-dir worldfoundry/data/models/catalog \
--requests-path tmp/requests.jsonl \
--output-dir tmp/model_eval/astra \
--metric artifact_count \
--jsonCompatibility & versions
Manifest-backed recipe
- Integration
- integrated
- Runner evidence
- verified
- Environment
- worldfoundry-unified-cu128
- Python
- 3.11
- CUDA
- CUDA 12.8
- PyTorch
- torch
- Source revision
- —
- Checkpoint revision
- aef5d6d75a994fa67f3f05cf17f5824a53b1c52f
- Runtime profile
- astra
- Pipeline binding
- astra
- Runner
- worldfoundry.pipeline
- Pipeline target
- worldfoundry.pipelines.kling.pipeline_astra:AstraPipeline
- Backend stage
- in_tree_runtime
- Runtime status
- checkpoint_verified
- Driver status
- compatible_unified_default
- Environment kind
- Unified 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 astraPackage constraints 12
torchtorchvisiondiffuserstransformershuggingface_hubsafetensorseinopsimageioimageio-ffmpegpillownumpytqdm
Conda packages 3
python=3.11pipffmpeg
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 astra --check-local --jsonEvanEternal/Astra
- Revision
- —
- License
- —
- Gated
- —
- Private
- —
Wan-AI/Wan2.1-T2V-1.3B
- Revision
- —
- License
- —
- Gated
- —
- Private
- —
EvanEternal/Astra
- Revision
- aef5d6d75a994fa67f3f05cf17f5824a53b1c52f
- License
- —
- Gated
- —
- Private
- —
Wan-AI/Wan2.1-T2V-1.3B
- Revision
- 37ec512624d61f7aa208f7ea8140a131f93afc9a
- 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 astra \
--model-runner worldfoundry:pipeline \
--model-manifest-dir worldfoundry/data/models/catalog \
--requests-path tmp/requests.jsonl \
--output-dir tmp/model_eval/astra \
--metric artifact_count \
--jsonInput contract
| Field | Recorded contract |
|---|---|
prompt | Required |
image | Required |
video | Optional |
actions | camera_direction |
Artifact contract
| Artifact kind | Filename / path |
|---|---|
generated_video | astra.mp4 |
Evidence & sources
Catalog integration, native-demo parity, and runner parity are independent records.
- Integration
- integrated
- Runner evidence
- verified
- Native demo evidence
- Not recorded
- Validation imports
- torch, torchvision, huggingface_hub, safetensors, einops, imageio, PIL, worldfoundry.pipelines.kling.pipeline_astra, worldfoundry.synthesis.visual_generation.kling.astra_synthesis
Recipe provenance
worldfoundry/data/models/catalog/world_models/astra.yamlProjecthttps://eternalevan.github.io/Astra-project/Paperhttps://arxiv.org/abs/2512.08931GitHubhttps://github.com/EternalEvan/AstraHugging Facehttps://huggingface.co/EvanEternal/Astra



