OpenVLA-OFT
Robotics modelOptimized Fine-Tuning (OFT) recipe and five MIT-licensed 7.54B-parameter LIBERO checkpoints built on OpenVLA-7B, replacing autoregressive token decoding with parallel continuous L1-regression action chunking, multi-camera inputs, proprioception, and optional FiLM language conditioning. [1] [2] [3] [4]
Setup & usage →CheckpointsModel hub / homepageGetting started
- For
- VLA researchers benchmarking on LIBERO and ALOHA; Robotics practitioners fine-tuning OpenVLA-7B with LoRA on custom RLDS datasets [1] [3] [5]
- Parameters
- 7,541,237,184 (~7.54B BF16 backbone params per HF safetensors metadata, plus lightweight MLP action head and… [4]
- Embodiments
- Franka Panda (LIBERO simulation), ALOHA bimanual manipulator (real robot, 14-DoF), WidowX (Bridge constants defined… [3] [5] [6]
- Source rights
- Named licenses are documented by part. Review licenses
Source researched · Not tested by us
Sources checked · Research details
Licenses & related hardware
Overview
OpenVLA-OFT (moojink/openvla-oft, arXiv:2502.19645) is an MIT-licensed fine-tuning recipe and checkpoint family built on openvla/openvla-7b. Instead of decoding discretized action tokens autoregressively one step at a time, OpenVLA-OFT adapts the 7.54B-parameter Prismatic VLM backbone with LoRA (r=32), parallel action-chunk decoding, a continuous MLP L1-regression head (or optional diffusion head), a proprioceptive state projector, multi-camera RGB inputs, and optional FiLM language conditioning (OFT+). [1] [3] [4] [5]
Model card summary
- Architecture
OpenVLA-7B Prismatic VLM backbone (DINOv2 + SigLIP vision encoders and Llama-2 7B) adapted with parallel decoding, continuous MLP L1-regression action head (or optional diffusion head), proprioception projector, multi-image inputs, and optional FiLM language-vision modulation (OFT+).
- Parameter count
7,541,237,184 (~7.54B BF16 backbone params per HF safetensors metadata, plus lightweight MLP action head and proprioception projector .pt modules)
- Input modalities
- rgbproprioceptionlanguage
2 RGB views (third-person + wrist) and 8-D proprioception for LIBERO; 3 RGB views (third-person + left/right wrist) and 14-D proprioception for ALOHA.
- Output modalities
- ee-pose-deltajoint-positionsgripper-command
7-DoF relative EEF pose + gripper for LIBERO and Bridge; 14-DoF absolute bimanual joint angles for ALOHA.
- Action space
Continuous parallel-decoded action chunks: 8-step x 7-DoF relative EEF deltas (BOUNDS_Q99) for LIBERO, 25-step x 14-DoF absolute joint angles (BOUNDS) for ALOHA, and 5-step x 7-DoF for Bridge per prismatic/vla/constants.py.
- Control frequency
25 Hz documented for ALOHA setup (with 25-step action chunks; 50 Hz supported with 50-step chunks); LIBERO closed-loop rate not pinned as a single Hz value in repo docs
- Trained / supported embodiments
- Franka Panda (LIBERO simulation)
- ALOHA bimanual manipulator (real robot, 14-DoF)
- WidowX (Bridge constants defined in codebase)
- Training datasets
- openvla/modified_libero_rlds (LIBERO-Spatial, LIBERO-Object, LIBERO-Goal, LIBERO-10 no-noops RLDS, ~10 GB)
- Custom ALOHA bimanual demonstration datasets converted to RLDS (user-collected; not hosted as prebuilt checkpoints)
Checkpoints
| Checkpoint | Format | Size | Access | Pinned revision |
|---|---|---|---|---|
| moojink/openvla-7b-oft-finetuned-libero-spatial | safetensors | 14.84 GB | open | 6d0231af0e48 |
| moojink/openvla-7b-oft-finetuned-libero-object | safetensors | 14.84 GB | open | 4c89574e1c53 |
| moojink/openvla-7b-oft-finetuned-libero-goal | safetensors | 14.84 GB | open | c2d0f9fbbd82 |
| moojink/openvla-7b-oft-finetuned-libero-10 | safetensors | 14.84 GB | open | 95220f9a3421 |
| moojink/openvla-7b-oft-finetuned-libero-spatial-object-goal-10 | safetensors | 14.84 GB | open | 638918f3d1c2 |
Each checkpoint repo contains BF16 safetensors backbone shards plus .pt files for the MLP action head and proprioception projector (14.84 GB total per repo).
Is this for you?
- Use cases
- Evaluating or fine-tuning continuous action-chunking VLA policies on LIBERO suites
- LoRA (r=32) fine-tuning of OpenVLA-7B with multi-camera RGB, proprioception, and FiLM conditioning on bimanual ALOHA tasks
- Serving fine-tuned VLA policies over FastAPI/uvicorn to a lightweight robot client
- Known limitations
- Upstream recommends evaluating on the same GPU model used for training because cross-GPU evaluation can cause substantial performance drops; merging LoRA weights on the target device may mitigate the drop
- Paper results depend on a custom transformers v4.40.1 fork (moojink/transformers-openvla-oft), Python 3.10.14, and PyTorch 2.2.0
- Published Hugging Face checkpoints cover LIBERO simulation suites only; real-world ALOHA checkpoints are not hosted on Hugging Face
- Checkpoints must be evaluated with center_crop=True to match the 90%-area random crop training augmentation
- Safety and limitations
- Evaluating a fine-tuned checkpoint on a different GPU architecture than the training GPU can substantially degrade performance unless LoRA adapters are merged on the target device.
- Base Llama-2 7B weights underlying OpenVLA-7B are subject to upstream OpenVLA / Llama-2 provenance even though the OpenVLA-OFT code and released checkpoints declare MIT.
- No safety interlocks or collision checks are included in the policy server or robot evaluation scripts.
Compute & fine-tuning
- Inference compute
- Minimum VRAM: 16 GB
- Documented hardware: NVIDIA A100 (upstream LIBERO paper evaluation)
- Precision:
bf16 - Edge / consumer capable: No
16 GB VRAM for LIBERO (18 GB for ALOHA) in bfloat16; exact per-step latency_ms not pinned in repo README.
- Fine-tuning
- Supported: Yes
- Methods: LoRA (r=32) fine-tuning with parallel decoding and L1 regression via vla-scripts/finetune.py; Optional diffusion action head (--use_diffusion True); Optional FiLM language conditioning (--use_film True, OFT+)
- Minimum VRAM: 25 GB
- Recipe: Upstream fine-tuning documentation
LIBERO.md reports ~25 GB VRAM at batch size 1 per GPU and ~62 GB at batch size 8 per GPU in bfloat16; paper reproduction uses 8 GPUs for 150K steps (50K for LIBERO-Goal).
- Host computer
Inference requires 1 GPU with ~16 GB VRAM for LIBERO and ~18 GB VRAM for ALOHA; fine-tuning requires 1–8 GPUs with 25–80 GB VRAM in bfloat16 (~25 GB at batch size 1 per GPU, ~62 GB at batch size 8 per GPU on LIBERO).
- Operating system
- Linux
Standard Linux conda, CUDA, flash-attn, and ROS/ALOHA client workflow documented.
- Dependencies
- Python 3.10 (3.10.14 tested)
- PyTorch 2.2.0
- moojink/transformers-openvla-oft (custom transformers v4.40.1 fork)
- flash-attn==2.5.5 (for training)
- openvla/openvla-7b (base weights for fine-tuning)
- uvicorn / fastapi / json-numpy (for ALOHA server deployment)
Setup & usage
Open the upstream getting-started guide [1] · Model card / documentation [1]
- Create a Python 3.10 conda environment and install PyTorch
- Clone moojink/openvla-oft and run pip install -e . (which installs the custom transformers-openvla-oft fork)
- Install flash-attn==2.5.5 for training and LIBERO or ALOHA requirements for evaluation
- Load a moojink/openvla-7b-oft-finetuned-libero-* checkpoint via get_vla, get_action_head, and get_proprio_projector, or run vla-scripts/finetune.py from openvla/openvla-7b
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 4-suite average (Spatial, Object, Goal, 10) | 97.1% average (4 task-specific checkpoints); 96.8% average (single combined 4-suite checkpoint) | Upstream-reported results on NVIDIA A100 averaged over 3 random seeds (500 trials/suite, custom transformers v4.40.1 fork); not independently reproduced by OSR. |
Project notes
Action chunking, multi-view inputs, and platform constants
In prismatic/vla/constants.py, the codebase defines three platform configurations: LIBERO uses 8-step action chunks over 7-DoF relative end-effector deltas (BOUNDS_Q99 normalization) with two RGB views (third-person plus wrist) and 8-D proprioception; ALOHA uses 25-step chunks over 14-DoF absolute bimanual joint angles (BOUNDS normalization, avoiding quantile clipping) at 25 Hz with three RGB views and 14-D proprioception; and Bridge defines 5-step 7-DoF chunks. Five ungated checkpoints are published on Hugging Face—four per-suite policies (libero-spatial, libero-object, libero-goal, libero-10) and one combined four-suite policy (libero-spatial-object-goal-10)—each totaling 14.84 GB across BF16 safetensors backbone shards (7,541,237,184 parameters) plus .pt weights for the MLP action head and proprioception projector. Real-world ALOHA checkpoints are not hosted on Hugging Face and must be fine-tuned from user demonstrations. [3] [4] [5] [6]
Compute requirements, deployment caveats, and upstream benchmarks
Upstream documents ~16 GB VRAM for LIBERO inference and ~18 GB VRAM for ALOHA inference in bfloat16, while LoRA fine-tuning requires ~25 GB VRAM at batch size 1 per GPU (~62 GB at batch size 8 per GPU on LIBERO). For LIBERO paper reproduction, upstream documents 8 GPUs for 150K steps (or the 50K-step checkpoint for LIBERO-Goal); for the documented real-world ALOHA put X into pot example, upstream instead specifies 8 GPUs and the 100K-step checkpoint. For real-robot ALOHA deployment, vla-scripts/deploy.py serves actions over FastAPI/uvicorn to a separate lightweight robot client environment. Three operational caveats are explicit in the upstream guides: evaluation requires --center_crop True to match training augmentations, LIBERO paper reproduction relies on a custom transformers v4.40.1 fork (moojink/transformers-openvla-oft) with Python 3.10.14 and PyTorch 2.2.0, and upstream recommends evaluating on the same GPU model used for training because cross-GPU evaluation can substantially degrade performance—merging LoRA weights into the base model on the target device via merge_lora_weights_and_save.py is noted as a possible mitigation rather than a guaranteed remedy. Upstream reports 97.1% average LIBERO success across the four task-specific checkpoints and 96.8% for the combined checkpoint (500 trials per suite, averaged over three seeds on NVIDIA A100). All benchmark figures are upstream-reported; Open Source Robotics has not run training, fine-tuning, inference, or robot evaluations for this profile. [1] [3] [5]
What we checked
None. Prepared from public upstream sources and Hugging Face API metadata without hands-on training, fine-tuning, or inference.
Updates
- Upstream last updated
- Sep 9, 2025 [1]
- Latest release
- Jun 17, 2025 [4]
- Profile updated
- Nov 21, 2026
- Sources last checked
- Nov 21, 2026
- Next review
- Feb 19, 2027
- Reviewed revision
- moojink/openvla-oft@e4287e94541f459edc4feabc4e181f537cd569a8; HF moojink/openvla-7b-oft-finetuned-libero-spatial@6d0231af0e48c5985f1ff86908f4674b84bc049b · combined hardware and software repository
Separate license matrix
MIT-licensed fine-tuning recipe and five ungated MIT-licensed 7.54B LIBERO VLA checkpoints built on OpenVLA-7B with public RLDS training datasets and reproducible LIBERO/ALOHA recipes.
| Layer | Availability / license |
|---|---|
| Code | available · MIT Verified MIT License in root LICENSE file. [1] [2] |
| Model weights | available · MIT All 5 Hugging Face LIBERO checkpoints declare license: mit in model card YAML frontmatter. [4] |
| Training datasets | available · MIT (LIBERO upstream benchmark; modified_libero_rlds hosted publicly on Hugging Face) LIBERO fine-tuning RLDS datasets (~10 GB) are publicly available; real-world ALOHA demonstrations are user-collected. [3] |
| Documentation | available · MIT [1] [2] |
Licensing and dataset availability
Both the moojink/openvla-oft codebase (verified in the root LICENSE file) and all five Hugging Face LIBERO checkpoints declare the MIT License. The LIBERO fine-tuning datasets (openvla/modified_libero_rlds, ~10 GB in RLDS format) are publicly hosted on Hugging Face, while ALOHA fine-tuning uses user-collected demonstrations converted via rlds_dataset_builder. This review verifies OpenVLA-OFT's MIT declarations for its code and LIBERO checkpoints, but does not resolve whether underlying openvla/openvla-7b (Llama-2 7B) base-model license obligations also apply; obtain and review the relevant OpenVLA and Llama-2 license texts before finalizing commercial compliance. [2] [3] [4] [5]
References
Each source links to the exact revision we read. “Used for” shows where this page relies on it.
- S001 moojink/openvla-oft GitHub repository metadata, README.md, and SETUP.md
Documents 16 GB LIBERO / 18 GB ALOHA inference VRAM, 1-8 GPU 27-80 GB training range, quickstart Python API, and arXiv:2502.19645.Revision, date & where cited
Moo Jin Kim (GitHub) · github.com · revision
e4287e9· GitHub API, README.md, SETUP.md at commit e4287e94541f (2025-09-09) · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Is this for you?, Compute & fine-tuning and 4 more - S002 moojink/openvla-oft root LICENSE at pinned commit
MIT License (Copyright 2025 Moo Jin Kim, Chelsea Finn, Percy Liang).Revision, date & where cited
Moo Jin Kim, Chelsea Finn, Percy Liang · github.com · revision
e4287e9· LICENSE at commit e4287e94541f · Accessed Nov 21, 2026
Used for: Separate license matrix - S003 moojink/openvla-oft LIBERO.md evaluation and fine-tuning guide
Reports 97.1% average LIBERO success (4 per-suite checkpoints) and 96.8% (combined checkpoint), ~25 GB / ~62 GB per-GPU VRAM at batch 1 / 8, custom transformers v4.40.1 fork, and GPU-mismatch warning.Revision, date & where cited
Moo Jin Kim (GitHub) · github.com · revision
e4287e9· LIBERO.md at commit e4287e94541f · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Checkpoints, Is this for you? and 6 more - S004 Hugging Face model cards and API metadata for 5 OpenVLA-OFT LIBERO checkpoints
All 5 checkpoints are ungated, license: mit, 7,541,237,184 BF16 safetensors parameters + .pt action head and proprio projector, 14.84 GB each.Revision, date & where cited
moojink (Hugging Face) · huggingface.co · HF API for 5 moojink/openvla-7b-oft-finetuned-libero-* repos (lastModified 2025-06-17) · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Checkpoints, Action chunking, multi-view inputs, and platform constants and 2 more - S005 moojink/openvla-oft ALOHA.md fine-tuning and server-client guide
Documents 25 Hz ALOHA control with 25-step action chunks, 3 camera views, FiLM language conditioning (OFT+), and FastAPI/uvicorn server-client deployment.Revision, date & where cited
Moo Jin Kim (GitHub) · github.com · revision
e4287e9· ALOHA.md at commit e4287e94541f · Accessed Nov 21, 2026
Used for: Overview, Model card summary, Is this for you?, Compute & fine-tuning and 3 more - S006 moojink/openvla-oft prismatic/vla/constants.py
Defines LIBERO (chunk 8, 7-DoF, 8-D proprio, BOUNDS_Q99), ALOHA (chunk 25, 14-DoF, 14-D proprio, BOUNDS), and BRIDGE (chunk 5, 7-DoF, 7-D proprio, BOUNDS_Q99).Revision, date & where cited
Moo Jin Kim (GitHub) · github.com · revision
e4287e9· prismatic/vla/constants.py at commit e4287e94541f · Accessed Nov 21, 2026
Used for: Model card summary, Action chunking, multi-view inputs, and platform constants
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 profile prepared from moojink/openvla-oft@e4287e94541f and five HF LIBERO checkpoints.