OhhO Train

Turn demonstrations into policies.

• Category: Intelligence
• Accent: cyan
• Live app: Open training console

The training engine behind the stack. Fine-tune VLA, imitation and reinforcement-learning policies from your OhhO Data episodes or in simulation — then export straight to Serve and Fleet.

Overview (hero)

OhhO Data collects the demonstrations; OhhO Serve runs the model — Train is the engine in between. Fine-tune Vision-Language-Action, imitation-learning and reinforcement-learning policies from your own episodes or in simulation, track every run, and export a deployment-ready checkpoint.

Highlights

• Fine-tune VLA, ACT, diffusion & RL policies
• Trains from OhhO Data or in simulation
• Built on the open OmniVLA engine
• Experiment tracking & sweeps (W&B)
• Exports to Serve + ONNX for Fleet OTA

What you get

• A policy is only as good as the loop that produced it. OhhO Train is the standardized training engine for embodied AI — the productized OmniVLA engine — so you fine-tune real models without stitching together a different toolchain for every method.
• Train covers the methods that matter: behavior cloning and VLA fine-tuning (SmolVLA, ACT, diffusion, OpenVLA) on the LeRobot datasets you record with OhhO Data, plus reinforcement learning in Isaac Lab with domain randomization for sim-to-real. A continual-learning loop can re-train as new episodes and tasks arrive, with replay, multi-objective rewards and AI-judged self-evaluation.
• Every run is tracked — losses, success rate, evaluation — with Weights & Biases sweeps to find good hyperparameters. When a checkpoint passes verification, Train exports it in the format the rest of the stack expects: a checkpoint OhhO Serve loads directly, or an ONNX policy OhhO Fleet ships over the air.

Features

• Many methods, one engine — Behavior cloning, VLA fine-tuning (SmolVLA / ACT / diffusion / OpenVLA), offline RL and on-policy RL — selected by config, not a rewrite.
• Trains from your data — Point Train at a LeRobot dataset from OhhO Data, or generate experience in Gazebo and Isaac Sim with domain randomization for sim-to-real.
• Continual learning — A post-training loop re-trains as new episodes and tasks arrive, with prioritized replay and outcome-stratified episodic memory.
• Rewards & self-evaluation — Multi-objective rewards (task, safety, efficiency, smoothness) plus vision rewards and AI judges score behavior, not just loss.
• Tracked & reproducible — Losses, success rate and eval stream to Weights & Biases; Bayesian sweeps search hyperparameters for you.
• Export-ready — Verified checkpoints export to OhhO Serve and to ONNX for OhhO Fleet OTA, with hardware-aware execution providers baked in.

How it works

1. Choose method & data — Pick a policy type and point Train at an OhhO Data dataset — or a simulation task.
2. Train & track — Launch the run; losses, success rate and eval stream to W&B in real time.
3. Verify — Best-of-N evaluation with hard safety and reachability checks gates what's allowed to ship.
4. Export & deploy — Push the checkpoint to OhhO Serve, or export ONNX for OhhO Fleet to roll out.

Specs

Plans

Recommended plan: Builder. Builder gives you cloud training with full experiment tracking — enough to fine-tune a first real policy. Teams running sweeps or standing up a continual-learning loop want Fleet; enterprises training on their own cluster choose Forge.

FAQ

Is the training engine open? Yes. Train is built on the open-source OmniVLA engine, so your training code and checkpoints are portable — you're never locked in.

Do I need real-robot data to start? No. You can train entirely in simulation with domain randomization, then fine-tune on real OhhO Data episodes for sim-to-real transfer.

Related products

• OhhO Data
• OhhO Serve
• OhhO Proof
• OhhO Market