Spirit-v1.5
Robotics modelVLA model implementation with RoboChallenge execution wrapper and separate base/fine-tuned checkpoints. [1] [3]
Setup & usage →CheckpointsModel hub / homepageGetting started
- For
- Robot-learning researchers familiar with PyTorch and benchmark-specific robot interfaces [1]
- Parameters
- 5,400,901,236 (~5.40B F32 per HF safetensors metadata) [31]
- Embodiments
- Bimanual 14-DoF manipulation robots (3-camera setup in config.json), RoboChallenge Table30 benchmark robots [1] [31]
- Source rights
- Some licenses are unresolved. Review licenses
Source researched · Not tested by us
Sources checked · Research details
Licenses & related hardware
- License
- MIT, Apache-2.0 · license not fully assessed for datasets · Details by part
Overview
Spirit-v1.5 (Spirit-AI-robotics/Spirit-v1.5) is a 5,400,901,236-parameter (~5.40B) Vision-Language-Action model from the Spirit AI Team pairing a Qwen/Qwen3-VL-4B-Instruct vision-language backbone with a 16-layer Diffusion Transformer (DiT) action head (dit_hidden_size: 1536, 32 attention heads, interleaved self-attention, and cross-attention dim 2560). Both the base model and a task-fine-tuned RoboChallenge checkpoint (Spirit-AI-robotics/Spirit-v1.5-for-RoboChallenge-move-objects-into-box) are published as ungated 20.12 GB FP32 safetensors files on Hugging Face Hub. [1] [30] [31] [32]
Model card summary
- Architecture
Qwen/Qwen3-VL-4B-Instruct vision-language backbone paired with a 16-layer Diffusion Transformer (DiT) action head (hidden_size: 1536, 32 heads, 10 denoising steps) using interleaved self-attention and cross-attention (dim 2560).
- Parameter count
5,400,901,236 (~5.40B F32 per HF safetensors metadata)
Exact tensor count verified via Hugging Face API.
- Input modalities
- rgbproprioceptionlanguage
Three RGB cameras (cam_high, cam_left_wrist, cam_right_wrist at 3x240x320), 14-D state (padded to 32), and text instruction.
- Output modalities
- joint-positionsgripper-command
14-D bimanual action chunks (padded to max_action_dim: 32).
- Action space
Continuous DiT action chunks with chunk_size: 60, n_action_steps: 60, num_steps: 10 denoising steps, and default 14-D action vector padded up to max_action_dim: 32 (MIN_MAX normalization).
- Trained / supported embodiments
- Bimanual 14-DoF manipulation robots (3-camera setup in config.json)
- RoboChallenge Table30 benchmark robots
- Training datasets
- Pretraining corpus not publicly released; fine-tuning example uses RoboChallenge/task_table30_move_objects_into_box
Checkpoints
| Checkpoint | Format | Size | Access | Pinned revision |
|---|---|---|---|---|
| Spirit-AI-robotics/Spirit-v1.5 | safetensors | 20.12 GB | open | 9886fa7552de |
| Spirit-AI-robotics/Spirit-v1.5-for-RoboChallenge-move-objects-into-box | safetensors | 20.12 GB | open | 03ed7624a1a9 |
Both ungated 20.12 GB FP32 safetensors checkpoints verified via Hugging Face API.
Is this for you?
- Use cases
- Fine-tuning the Spirit-v1.5 base VLA on custom bimanual/tabletop manipulation datasets via FSDP
- Reproducing RoboChallenge Table30 benchmark evaluations with task-fine-tuned checkpoints
- Known limitations
- Base-card Apache2 and code MIT are distinct declarations; backbone, data and task-checkpoint terms remain separate.
- Training recommendation8xA10080GB is upstream workload advice, not our requirement measurement.
- No weights downloaded, model run, hardware command or benchmark reproduction.
Editorial scope assessment from the cited implementation/documentation; no execution or compatibility test.
- Safety and limitations
- Repository README (Python 3.10+, torch 2.8.0, transformers 4.57.1, diffusers 0.35.2) and base model card (Python 3.11+, torch 2.9.1, transformers 4.57.3, diffusers 0.36.0) list different pinned dependency versions that must be reconciled before install.
- Base checkpoint and task-fine-tuned RoboChallenge checkpoint serve different workflows and are not interchangeable.
- Pretraining dataset is not publicly released; code is MIT while both inspected checkpoints declare Apache-2.0.
Compute & fine-tuning
- Inference compute
- Documented hardware: NVIDIA A100 80GB (upstream test environment, not a measured minimum)
- Precision:
fp32 - Edge / consumer capable: No
20.12 GB FP32 safetensors footprint; exact minimum VRAM and latency_ms not measured here.
- Fine-tuning
- Supported: Yes
- Methods: Multi-GPU PyTorch FSDP fine-tuning via scripts/run_finetune.sh (recommended 8x A100 80GB)
- Recipe: Upstream fine-tuning documentation
- Host computer
NVIDIA A100 80GB GPU tested upstream for inference; 8x A100 80GB recommended upstream for FSDP fine-tuning.
- Operating system
- Linux
PyTorch CUDA / torchrun Linux workflow documented.
[1] · Editorial assessment
- Dependencies
- Repository README Python3.10+ versus base-card recommended Python3.11+
- Source reports inference testing on NVIDIA A10080GB, not a measured minimum
- Separate requirements-base.txt and requirements-train.txt; flash-attn needs matching PyTorch/CUDA
- Skills
- Python
- PyTorch
- FSDP / torchrun
- CUDA / flash-attn
Inferred from documented PyTorch, FSDP, and flash-attn setup.
[1] · Editorial assessment
Setup & usage
Open the upstream getting-started guide [1]
- Choose inference-only or training dependencies and a named base/fine-tuned checkpoint; follow the selected runtime configuration rather than treating all checkpoints as the same robot policy.
- Preserve exact source/dependency/model versions and evaluate the selected workflow before connecting physical hardware or reusing data.
Editorial scope assessment from the cited implementation/documentation; no execution or compatibility test.
Upstream-reported benchmarks
Upstream-reported only: Open Source Robotics has not run training, inference, or hardware evaluations for these benchmarks.
| Benchmark | Upstream-reported score | Scope & caveats |
|---|---|---|
| RoboChallenge Table30 | Ranked #1 on RoboChallenge Table30 leaderboard as of 2026-01-11 (upstream README claim) | Upstream-reported ranking in README.md and Spirit AI blog; not independently verified by OSR. |
Project notes
Action space and checkpoint scope
At the pinned base revision (9886fa7552de), config.json defines three RGB camera inputs (cam_high, cam_left_wrist, cam_right_wrist at 3x240x320), a 14-D proprioceptive state (max_state_dim: 32), and text instructions, outputting continuous DiT action chunks with chunk_size: 60, n_action_steps: 60, num_steps: 10 denoising steps, and default 14-D bimanual actions (max_action_dim: 32, MIN_MAX normalization). Choosing the base checkpoint for FSDP fine-tuning (scripts/run_finetune.sh) is distinct from running the RoboChallenge evaluation wrapper (scripts/run_robochallenge.sh): fine-tuning installs requirements-train.txt on top of requirements-base.txt and requires DATA_ROOT plus PRETRAINED_PATH pointing to a checkpoint directory containing both model.safetensors and config.json, whereas RoboChallenge evaluation uses requirements-base.txt and requires a TASK_NAME defined in robochallenge/runner/task_info.py, ROBOCHALLENGE_JOB_ID, USER_TOKEN, and CKPT_PATH pointing to the directory containing model.safetensors. [1] [31] [32]
Environment discrepancy, compute, and upstream-reported benchmarks
The repository README specifies Python 3.10+ with torch 2.8.0, transformers 4.57.1, and diffusers 0.35.2 (plus optional flash-attn 2.8.3 for training), whereas the base Hugging Face model card recommends Python 3.11+ with torch 2.9.1, transformers 4.57.3, and diffusers 0.36.0; those pinned versions must be reconciled before installation. Upstream reports inference testing on an NVIDIA A100 80 GB and recommends 8x A100 80 GB for FSDP fine-tuning, and states that Spirit-v1.5 ranked #1 on the RoboChallenge Table30 leaderboard as of January 11, 2026. Minimum VRAM, inference latency, and control frequency are not established in the reviewed evidence—the tested A100 80 GB GPU is not a minimum requirement, and the 60-step chunk length alone does not establish execution duration. All benchmark rankings and GPU figures are upstream-reported; Open Source Robotics has not downloaded weights or run training, fine-tuning, inference, or robot evaluations. [1] [30] [31]
What we checked
Source research only. Selected implementation/model text, repository tree and license declarations were inspected. No upstream compilation, dependency installation, simulator, network/USB/device integration, data collection, benchmark or physical robot test was performed. Source files are linked rather than redistributed; external services and paid products are not endorsed.
Not yet established
- Documented closed-loop control frequency (Hz)
Updates
Separate license matrix
Distinct robot-facing model contribution with actual implementation/model content and a bounded documented next step. Review covers source artifacts, not operational performance or a complete tested deployment.
| Layer | Availability / license |
|---|---|
| Code | available · MIT Declaration for the scoped source implementation/documentation; dependencies, external services, model/data assets and inherited file-specific terms require separate checks. No blanket redistribution or operational clearance. [1] [2] [3] |
| Model weights | available · Apache-2.0 Both inspected Hugging Face checkpoints (base Spirit-v1.5 and RoboChallenge-move-objects-into-box) are ungated and declare Apache-2.0; weight binaries not downloaded or executed. [30] [31] [32] |
| Training datasets | available · License not fully assessed Pretraining corpus is not publicly released; upstream fine-tuning guide references public RoboChallenge task datasets whose licenses remain separate. [1] |
| Documentation | available · MIT Declaration for the scoped source implementation/documentation; dependencies, external services, model/data assets and inherited file-specific terms require separate checks. No blanket redistribution or operational clearance. [1] [2] [3] |
Separate code, weight, and dataset rights
The spirit-ai-team/spirit-v1.5 codebase is MIT-licensed, while both inspected Hugging Face checkpoints declare Apache-2.0. The pretraining dataset is not publicly released; upstream's fine-tuning guide references public RoboChallenge task datasets (RoboChallenge/task_table30_move_objects_into_box), whose licenses remain separate. [1] [2] [30] [31] [32]
References
Each source links to the exact revision we read. “Used for” shows where this page relies on it.
- S001 Project README
Revision, date & where cited
spirit-ai-team/spirit-v1.5 contributors · github.com · revision
6d67377· Project README at · Accessed Nov 1, 2026
Used for: Overview, Model card summary, Is this for you?, Compute & fine-tuning and 6 more - S002 License declaration
Revision, date & where cited
spirit-ai-team/spirit-v1.5 contributors · github.com · revision
6d67377· License declaration at · Accessed Nov 1, 2026
Used for: Is this for you?, Compute & fine-tuning, Setup & usage, Separate license matrix - S003 Selected implementation excerpt
Revision, date & where cited
spirit-ai-team/spirit-v1.5 contributors · github.com · revision
6d67377· Selected implementation excerpt at · Accessed Nov 1, 2026
Used for: Model card summary, Is this for you?, Compute & fine-tuning, Setup & usage and 1 more - S004 Repository artifact tree
Revision, date & where cited
spirit-ai-team/spirit-v1.5 contributors · github.com · revision
6d67377· Repository artifact tree at · Accessed Nov 1, 2026
Used for: Is this for you?, Compute & fine-tuning, Setup & usage - S030 Published checkpoint card and metadata
Revision, date & where cited
spirit-ai-team/spirit-v1.5 contributors · huggingface.co · Published checkpoint card and metadata at · Accessed Nov 1, 2026
Used for: Overview, Model card summary, Is this for you?, Compute & fine-tuning and 3 more - S031 Hugging Face API metadata and config.json: Spirit-AI-robotics/Spirit-v1.5
5,400,901,236 F32 params (20.12 GB safetensors), ungated, license: apache-2.0, backbone: Qwen/Qwen3-VL-4B-Instruct, 16-layer DiT (hidden_size 1536, 32 heads, cross_attn 2560), chunk_size: 60, n_action_steps: 60, num_steps: 10, 3 cameras (3x240x320), 14-D state/action padded to 32, lastModified 2026-01-11.Revision, date & where cited
Hugging Face (Spirit-AI-robotics) · huggingface.co · HF API & config.json at revision · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Checkpoints, Is this for you? and 4 more - S032 Hugging Face API metadata: Spirit-AI-robotics/Spirit-v1.5-for-RoboChallenge-move-objects-into-box
5,400,901,236 F32 params (20.12 GB safetensors), ungated, license: apache-2.0, base_model: Spirit-AI-robotics/Spirit-v1.5, lastModified 2026-01-11.Revision, date & where cited
Hugging Face (Spirit-AI-robotics) · huggingface.co · HF API metadata at revision · Accessed Nov 21, 2026
Used for: Overview, Checkpoints, Is this for you?, Action space and checkpoint scope and 1 more
Correction guidance · Discussion is not enabled · Profile format 0.8.0-draft.1
Change history
2026-11-21 — Migrated 0.7.0-draft.5 -> 0.8.0-draft.1 (kind=model); added verified HF API and config.json metadata (S031, S032) while preserving historical source dates.