X-VLA
Robotics model880M-parameter (~0.9B) soft-prompted flow-matching Vision-Language-Action model (ICLR 2026) built on Florence-2-large with embodiment-specific learnable soft prompts, Apache-2.0 code and weights, LoRA fine-tuning, and native LeRobot integration. [1] [2] [4] [5]
Setup & usage →CheckpointsModel hub / homepageGetting started
- For
- Robotics researchers; Cross-embodiment VLA engineers; LeRobot practitioners [4]
- Embodiments
- Franka Panda (LIBERO, CALVIN, VLABench), WidowX 250 (BridgeData V2 / SimplerEnv), AgileX (RoboTwin 2.0, SoftFold… [4]
- Source rights
- Named licenses are documented by part. Review licenses
Source researched · Not tested by us
Sources checked · Research details
Licenses & related hardware
- Ecosystem
- LeRobot
Overview
X-VLA (2toINF/X-VLA-Pt and native LeRobot port lerobot/xvla-base) is an 879,738,545-parameter (~0.88B / 0.9B) flow-matching Vision-Language-Action model (ICLR 2026) built on an encoder-only microsoft/Florence-2-large vision-language backbone paired with a 24-layer SoftPromptedTransformer action denoiser. It supports full fine-tuning and LoRA (PEFT) adapters alongside a FastAPI inference server/client. [1] [2] [3] [4]
Model card summary
- Architecture
Florence-2-large encoder-only vision-language backbone paired with a 24-layer SoftPromptedTransformer flow-matching action denoiser (hidden_size: 1024, 16 heads, 32 learnable soft prompt tokens across up to 30 domains) and an Action Hub.
Verified from HF model card and config.json.
- Base model
microsoft/Florence-2-large
- Parameter count
879,738,545 (~0.88B / '0.9B' F32 parameters per safetensors metadata)
Exact identical parameter count verified on both 2toINF/X-VLA-Pt and lerobot/xvla-base.
- Input modalities
- rgblanguageproprioception
Primary/multi-view RGB images, English language instruction, and proprioceptive state (use_proprio: true).
- Output modalities
- ee-pose-absee-pose-deltajoint-positionsgripper-command
Configurable per-embodiment Action Hub (default action_mode: ee6d in X-VLA-Pt config.json).
- Action space
Flow-matching continuous action chunks (num_actions: 30, default action_mode: ee6d with 6D end-effector rotation representation) managed through embodiment-specific Action Hub masks and soft prompts (num_domains: 30).
- Trained / supported embodiments
- Franka Panda (LIBERO, CALVIN, VLABench)
- WidowX 250 (BridgeData V2 / SimplerEnv)
- AgileX (RoboTwin 2.0, SoftFold dual-arm)
- AgiBot-G1 (IROS 2025 AgiBot World Challenge)
- Google Robot (SimplerEnv)
- Training datasets
- bridgedata-v2
- Heterogeneous cross-embodiment pretraining mixture (LIBERO, CALVIN, VLABench, RoboTwin 2.0, Soft-Fold, Google Robot, AgiBot World)
Checkpoints
| Checkpoint | Format | Size | Access | Pinned revision |
|---|---|---|---|---|
| 2toINF/X-VLA-Pt (Foundation pretrained) | safetensors | 3.52 GB | open | c1c4a64a7e03 |
| lerobot/xvla-base (LeRobot-native port) | safetensors | 3.52 GB | open | cdb7964e4fe8 |
Per-checkpoint repository storage is ~3.52 decimal GB (3.28 GiB; 3,519,073,692 bytes in S005 metadata); exact individual safetensors file size is not separated from repository storage in S005.
Is this for you?
- Use cases
- Cross-embodiment VLA fine-tuning (full or LoRA PEFT)
- Client-server HTTP policy deployment on real robots
- Deformable object manipulation and dual-arm coordination
- Known limitations
- Authors report an ~1% performance drop after converting original checkpoints to Hugging Face format (under active investigation per README)
- Direct Transformers loading uses trust_remote_code=True
- Safety and limitations
- Upstream reports an approximately 1% performance drop across datasets after converting checkpoints to Hugging Face format (metric unspecified; cause under investigation at reviewed revision; parity with lerobot/xvla-base not established).
- Loading 2toINF checkpoints via AutoModel requires trust_remote_code=True (native lerobot/xvla-base avoids custom Hub code); client inference requires matching domain_id and training-space proprioception normalization.
Compute & fine-tuning
- Inference compute
- Precision:
fp32 - Edge / consumer capable: Yes
3.28 GB F32 safetensors weights; README notes suitability for GPUs, SLURM clusters, or edge devices; exact min_vram_gb and latency_ms not specified.
- Precision:
- Fine-tuning
- Supported: Yes
- Methods: full fine-tune; LoRA (PEFT adapters)
- Recipe: Upstream fine-tuning documentation
- Host computer
3.28 GB F32 checkpoint designed for GPU servers, SLURM clusters, or edge devices via decoupled client-server inference.
- Operating system
- Linux
Inferred from documented Linux/CUDA Python environment and shell instructions.
[4] · Editorial assessment
- Dependencies
- Python 3.10
- PyTorch
- Transformers
- FastAPI / json_numpy
- LeRobot (optional native integration)
- Skills
- Python
- PyTorch
- Hugging Face Transformers / PEFT LoRA
- FastAPI client-server robotics setup
Inferred from documented Python/PyTorch/CLI setup and training dependencies.
[4] · Editorial assessment
Setup & usage
Open the upstream getting-started guide [4] · Model card / documentation [1] [2]
- Clone 2toinf/X-VLA and create Python 3.10 conda environment
- Load 2toINF/X-VLA-Pt (or embodiment checkpoint) via AutoModel.from_pretrained(..., trust_remote_code=True) or use native LeRobot integration (lerobot/xvla-base)
- Start FastAPI server (model.run(processor, host='0.0.0.0', port=8000)) and query POST /act from robot client
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 |
|---|---|---|
| LIBERO / CALVIN ABC_D / SimplerEnv WidowX / RoboTwin 2.0 / IROS 2025 AgiBot World Challenge | LIBERO 98.1%, CALVIN ABC_D 4.43, SimplerEnv WidowX 95.8%, RoboTwin 2.0 70%, VLABench 51.1, IROS 2025 AgiBot World Challenge 1st Place | Upstream-reported benchmark scores from 2toinf/X-VLA README (fine-tuned checkpoints); not independently verified by OSR. |
Project notes
Checkpoint selection and soft-prompted domain adaptation
Upstream separates the foundation pretrained checkpoint (2toINF/X-VLA-Pt), the native LeRobot port (lerobot/xvla-base), and embodiment-specific fine-tuned checkpoints; choose a foundation checkpoint when fine-tuning on a new robot or task, and use an embodiment-specific checkpoint only when your setup matches that domain. At pinned Hub revision c1c4a64a7e03, config.json specifies hidden_size: 1024, depth: 24, num_heads: 16, num_actions: 30, action_mode: ee6d (6D end-effector rotation), use_proprio: true, num_domains: 30, and len_soft_prompts: 32. Each robot domain uses learnable soft prompts and Action Hub masks; the documented inference client supplies a domain_id, and proprioceptive inputs must match training-space normalization—selecting a domain_id alone does not adapt the model to an unseen robot. [2] [3] [4] [5]
Upstream reports an approximately 1% performance drop across datasets after converting checkpoints to Hugging Face format (the metric is unspecified, and upstream noted the cause was still under investigation at the reviewed revision). Loading 2toINF/X-VLA-Pt via AutoModel requires trust_remote_code=True, whereas lerobot/xvla-base uses native LeRobot code; the supplied evidence does not establish conversion parity or whether the reported conversion drop applies to lerobot/xvla-base. [4] [5]
Footprint and upstream-reported benchmarks
Reported repository storage is approximately 3.52 decimal GB (3.28 GiB; 3,519,073,692 bytes in S005) per ungated FP32 checkpoint (metadata reports total repository storage rather than isolating exact individual weight-file size). Upstream does not pin a single minimum VRAM or closed-loop latency scalar in the README. The repository reports fine-tuned benchmark scores including LIBERO (98.1%), CALVIN ABC_D (4.43), SimplerEnv WidowX (95.8%), RoboTwin 2.0 (70%), VLABench (51.1), and 1st place in the IROS 2025 AgiBot World Challenge. All scores are upstream-reported; Open Source Robotics has not run training, fine-tuning, inference, or hardware evaluations for X-VLA. [2] [4] [5]
What we checked
None. Prepared from public upstream sources without hands-on hardware execution.
Not yet established
- Documented closed-loop control frequency (Hz)
Updates
- Upstream last updated
- Oct 20, 2026 [1]
- Latest release
- Mar 2, 2026 [5]
- Profile updated
- Nov 21, 2026
- Sources last checked
- Nov 21, 2026
- Next review
- May 21, 2027
- Reviewed revision
- 2toinf/X-VLA@6c4ecfb8be458ae7c0c96094e4b0add4174f4fe2; HF 2toINF/X-VLA-Pt@c1c4a64a7e03ac5b95c468bf1578f3d03651b53b; HF lerobot/xvla-base@cdb7964e4fe842935d671bfab5a5ebe00a96648c · software repository
Separate license matrix
Clean Apache-2.0 0.88B soft-prompted flow-matching VLA with ungated safetensors weights, LoRA support, and native LeRobot integration.
| Layer | Availability / license |
|---|---|
| Code | available · Apache-2.0 Apache License Version 2.0 verified in root LICENSE. [1] [6] |
| Model weights | available · Apache-2.0 Both 2toINF/X-VLA-Pt and lerobot/xvla-base declare license: apache-2.0; upstream license terms of base model microsoft/Florence-2-large were not excerpted in this review and require independent verification. [2] [5] |
| Training datasets | available · Mixed upstream (heterogeneous cross-embodiment and benchmark datasets; individual upstream dataset terms not excerpted in this review) Pretraining and benchmark fine-tuning datasets carry separate upstream licenses. [4] |
| Documentation | available · Apache-2.0 [4] [6] |
Separate code, weights, and pretraining dataset licenses
Both 2toinf/X-VLA (verified in root LICENSE) and both Hugging Face checkpoints (2toINF/X-VLA-Pt and lerobot/xvla-base) carry Apache-2.0 licenses. In contrast, the upstream license terms of the microsoft/Florence-2-large base backbone and the heterogeneous pretraining and benchmark datasets (such as BridgeData V2, LIBERO, CALVIN, RoboTwin, and AgiBot World) were not excerpted in this review and require independent verification. [1] [2] [4] [5] [6]
References
Each source links to the exact revision we read. “Used for” shows where this page relies on it.
- S001 2toinf/X-VLA GitHub repository metadata and HEAD commit
Default branch main HEAD is 6c4ecfb8be458ae7c0c96094e4b0add4174f4fe2 (2026-10-20T05:16:33Z); SPDX license Apache-2.0.Revision, date & where cited
2toinf (GitHub) · github.com · revision
6c4ecfb· GitHub API repos/2toinf/X-VLA & commits/main · Accessed Nov 21, 2026
Used for: Overview, Setup & usage, What we checked, Separate license matrix - S002 Hugging Face model card and API metadata: 2toINF/X-VLA-Pt
879,738,545 F32 parameters, 3.28 GB safetensors, ungated, license: apache-2.0, base_model: microsoft/Florence-2-large.Revision, date & where cited
2toINF (Hugging Face) · huggingface.co · HF API & README.md at revision c1c4a64a7e03 · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Checkpoints, Is this for you? and 5 more - S003 2toINF/X-VLA-Pt config.json at pinned revision
action_mode: ee6d, use_proprio: true, num_actions: 30, hidden_size: 1024, depth: 24, num_heads: 16, num_domains: 30, len_soft_prompts: 32.Revision, date & where cited
2toINF (Hugging Face) · huggingface.co · config.json at revision c1c4a64a7e03 · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Checkpoint selection and soft-prompted domain adaptation - S004 2toinf/X-VLA README.md at pinned commit
Documents ICLR 2026 acceptance, LeRobot integration, IROS 2025 AgiBot World Challenge 1st place, LoRA adapters, benchmark table, and FastAPI server/client.Revision, date & where cited
2toinf (GitHub) · github.com · revision
6c4ecfb· README.md at commit 6c4ecfb8be45 · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Is this for you?, Compute & fine-tuning and 5 more - S005 Hugging Face API metadata: lerobot/xvla-base
879,738,545 F32 parameters, 3.28 GB safetensors, ungated, license: apache-2.0, lastModified 2026-03-02.Revision, date & where cited
Hugging Face (lerobot) · huggingface.co · HF API /api/models/lerobot/xvla-base · Accessed Nov 21, 2026
Used for: Model card summary, Checkpoints, Is this for you?, Checkpoint selection and soft-prompted domain adaptation and 3 more - S006 2toinf/X-VLA LICENSE at pinned commit
Revision, date & where cited
github.com · revision
6c4ecfb· LICENSE at commit 6c4ecfb8be45 · Accessed Nov 21, 2026
Used for: Separate license matrix
Correction guidance · Discussion is not enabled · Profile format 0.8.0-draft.1
Change history
2026-11-21 — Initial 0.8.0-draft.1 model draft prepared from pinned GitHub commit 6c4ecfb8be45 and HF revisions c1c4a64a7e03 / cdb7964e4fe8.