Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline)

RL Training

The whole system is an asynchronous loop of three independent components that communicate only through HuggingFace Hub (Figure 2 ): Training worker (one GPU machine): trains continuously, recomputes advantages across all rollout datasets before each iteration, uploads checkpoints.

Rollout workers (any number of machines): pulls the latest checkpoint from the Hub, runs 3–5 parallel Isaac Sim instances each, uploads collected episodes with the values predicted during collection.

Data Collection

Simulation speed is the bottleneck of the whole system — an episode takes about 30 s in Isaac Sim — so the collection side is built around squeezing useful episodes out of every sim-hour. The collection stack went through a lot of engineering optimization; I won’t cover it in detail here, but the full implementation is released with the code.

Policy Architecture

I start from the architecture my team built for the BEHAVIOR-1K Challenge [ 7 ] , which itself extends $\pi_{0.5}$ [ 8 ] . The base stack: SigLIP-So400m/14 image encoder [ 15 ] $\to$ Gemma-2B prefix transformer [ 16 ] (images + state + auxiliary tokens) $\to$ Gemma-300M action expert that generates a 30-step action chunk (1 s at 30 Hz, 12-dim joint deltas) via flow matching [ 17 ] , attending to the prefix KV cache.

Three RGB cameras (top, left wrist, right wrist), all resized to 224$\times$224. The vision backbone is frozen.

Auxiliary Prediction Heads

I use two groups of auxiliary prediction heads: current-frame heads (§5.1) that include the success and completion predictions used as a value function, and future-prediction heads (§5.2) that act as a Q-function and a cheap quasi-world model, predicting what the scene will look like 30 frames ahead.

All current-frame prediction heads are linear probes off a single learned current query token (1$\times$2048) placed in the image attention group (§4.1): they see the images and nothing else, and they all share one VLM forward pass.

Reward Design and Advantage Computation

Reward engineering is one of the most important steps in RL training. While advantage estimation given a reward is usually theoretically grounded, the choice of reward itself is partly an “art” built on heuristics and experiments. This section explains how I arrived at the final advantage function, step by step, including the subjective choices.

The competition metric is a binary episode success — too sparse for efficient RL: correct actions early in the episode get almost no signal, and an easy success is indistinguishable from a hard one.

Inference-Time Optimization

The same model checkpoint can achieve different success rates depending on how it is run: how much of each predicted chunk is actually executed, how strongly the advantage conditioning is amplified, how many candidate chunks are drawn and re-ranked. None of these touch the weights — they are all decided at deployment, and they make a large difference.

This section covers every knob applied at inference and how I tuned them per garment type.

Online Round Results

The online round of the LeHome Challenge 2026 is the simulation track: the policy runs in Isaac Sim against the four garment types, each episode is scored as a binary fold success, and teams are ranked on a public leaderboard by overall success rate. The leaderboard set is 20 garments per type (80 in total): the 10 seen garments the organizers released training data for, and 10 unseen ones (§1.1).

Sim-to-Real Transfer (Final Round)

The competition had two rounds. Everything up to this point — the policy, the RL pipeline, the inference tuning, the 1st-place online result of Section 8 — was the simulation round. The final round was sim-to-real: take a policy trained in Isaac Sim and make it fold real garments on a real bimanual robot. I had roughly three months for the simulation round and a bit more than one week for sim-to-real.

Sources and Demonstrations

Figures are reproduced from Larchenko See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.