OpenVLA OFT Fine Tuning and Evaluation
Skill Verified ActiveFine-tunes and evaluates OpenVLA-OFT and OpenVLA-OFT+ policies for robot action generation with continuous action heads, LoRA adaptation, and FiLM conditioning on LIBERO simulation and ALOHA real-world setups. Use when reproducing OpenVLA-OFT paper results, training custom VLA action heads (L1 or diffusion), deploying server-client inference for ALOHA, or debugging normalization, LoRA merge, and cross-GPU issues.
Enables researchers and engineers to reproduce OpenVLA-OFT paper results, train custom VLA action heads, and deploy server-client inference for robotics applications.
Features
- Fine-tuning and evaluation of OpenVLA-OFT and OFT+
- LoRA adaptation for efficient fine-tuning
- Continuous action heads (L1 regression or diffusion)
- FiLM conditioning for enhanced language grounding (OFT+)
- Support for LIBERO simulation and ALOHA real-world setups
- Server-client deployment for ALOHA inference
- Detailed troubleshooting for common issues and invariants
Use Cases
- Reproducing OpenVLA-OFT paper results
- Training custom VLA action heads (L1 or diffusion)
- Deploying server-client inference for ALOHA robots
- Debugging normalization, LoRA merge, and cross-GPU issues
Non-Goals
- General LLM fine-tuning without robot action heads
- Fine-tuning other VLA architectures (e.g., pi0/pi0.5 models)
- Using the NVIDIA Cosmos Policy stack
Practices
- Model Fine-Tuning
- Robot Action Generation
- Simulation Environments
- Real-World Deployment
- Model Evaluation
- Troubleshooting
Prerequisites
- Python 3.10+
- PyTorch 2.2.0+
- Transformers >=4.40.0
- PEFT ==0.11.1
- Specific GPU VRAM requirements (see SKILL.md)
Installation
First, add the marketplace
/plugin marketplace add Orchestra-Research/AI-Research-SKILLs/plugin install AI-Research-SKILLs@ai-research-skillsQuality Score
VerifiedTrust Signals
Similar Extensions
Peft Fine Tuning
99Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.
OpenPI Fine Tuning and Serving
98Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robot policy inference across ALOHA, DROID, and LIBERO environments. Use when adapting pi0 models to custom datasets, converting JAX checkpoints to PyTorch, running policy inference servers, or debugging norm stats and GPU memory issues.
Unsloth
98Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization
Implementing Llms Litgpt
98Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean model implementations, educational understanding of architectures, or production fine-tuning with LoRA/QLoRA. Single-file implementations, no abstraction layers.
Fine Tuning Expert
98Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters for fine-tuning runs, adapter training, transfer learning, finetuning with Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO, or quantizing and deploying fine-tuned models. Trigger terms include: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model.
Axolotl Fine Tuning Skill
98Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support