KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#typescript #ai-agents #cli #claude-code #codex #developer-tools #ai #dsh-plugin #ai-agent #mcp #deepseek-harness #windows

WEAVER

★ 38 Open GitHub ↗

WEAVER is an efficient, consistent and high-fidelity world model enabling state-of-the-art robot policy evaluation, improvement, and faster test-time planning.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# WEAVER, Better, Faster, Longer: An Effective World Model for Robotic Manipulation

<a href="https://arnavkj1995.github.io/">Arnav Kumar Jain</a>1,2,*,
 <a href="https://yilin-wu98.github.io/">Yilin Wu</a>3,*,
 <a href="https://brosa.ca/">Jesse Farebrother</a>1,4,
 <a href="https://gokul.dev/">Gokul Swamy</a>3,
 <a href="https://www.cs.cmu.edu/~abajcsy/">Andrea Bajcsy</a>3

[![arXiv](https://img.shields.io/badge/arXiv-2506.05294-df2a2a.svg?style=for-the-badge&logo=arxiv)](https://arxiv.org/abs/2606.13672)
[![Website](https://img.shields.io/badge/Website-WEAVER-black?style=for-the-badge)](https://arnavkj1995.github.io/WEAVER/)
[![Summary](https://img.shields.io/badge/Summary-1DA1F2?logo=x&logoColor=white&labelColor=gray&style=for-the-badge)](https://x.com/arnavkj95/status/2066919523685380387?s=20)
[![Models](https://img.shields.io/badge/Models-%F0%9F%A4%97-yellow?style=for-the-badge)](https://huggingface.co/arnavkj1995/WEAVER)
[![Dataset](https://img.shields.io/badge/Dataset-%F0%9F%A4%97-yellow?style=for-the-badge)](https://huggingface.co/datasets/yilin-wu/droid_ood_data)
[![License](https://img.shields.io/badge/License-MIT-green?style=for-the-badge)](https://raw.githubusercontent.com/arnavkj1995/WEAVER/main/LICENSE)
---

We introduce WEAVER: a world model architecture that satisfies the three desiderata: (i) fidelity, (ii) consistency, and (iii) efficiency. WEAVER unlocks state-of-the-art performance across policy evaluation (ρ = 0.870 correlation with real-world success rate), policy improvement (real-world success rate improvement of 38% on top of the π0.5 robot foundation model), and test-time planning (real-world success rate improvement of 14% with a 5–10× speedup over prior WMs).

## 🛠️ Setup

Create a Python 3.11 environment and install dependencies with `uv`.

```bash
git clone --recurse-submodules https://github.com/arnavkj1995/WEAVER.git
cd WEAVER
uv venv --python 3.11
source .venv/bin/activate
uv sync
```

For optional logging and development dependencies:

```bash
uv sync --extra logging --extra dev
```

You can also run commands directly through `uv`:

```bash
uv run python -m weaver.generate_views --help
```

## Model Checkpoints and Datasets

Download all released checkpoints from [HuggingFace](https://huggingface.co/arnavkj1995/WEAVER):

```bash
hf download arnavkj1995/WEAVER \
  --local-dir checkpoints
```

This downloads the `WEAVER`, `WEAVER-FT`, and `WEAVER-ReFlow` folders. Each
folder contains `checkpoint.pt`, `config.yaml`, and `norm_stats_relabel.json`.

The OOD evaluation dataset used in this work is available at
[yilin-wu/droid_ood_data](https://huggingface.co/datasets/yilin-wu/droid_ood_data).

```bash
# Download the full dataset
git lfs install
git clone https://huggingface.co/datasets/yilin-wu/droid_ood_data
```

Or with the Python library:

```python
from huggingface_hub import snapshot_download

local_dir = snapshot_download(
    repo_id="yilin-wu/droid_ood_data",
    repo_type="dataset",
)
```

To download only annotations and metadata (without videos/latents):

```python
from huggingface_hub import snapshot_download

local_dir = snapshot_download(
    repo_id="yilin-wu/droid_ood_data",
    repo_type="dataset",
    allow_patterns=["annotations/**", "annotation_rewards/**", "norm_stats*.json"],
)
```

## 📁 Repository Structure

This repository implements the main WEAVER components: the latent flow world model, DROID dataloaders, reward and critic heads, rollout generation, and offline evaluation utilities.

```text
WEAVER
├── assets                           # README and release assets
├── datasets                         # DROID preprocessing and SD3 latent encoding utilities
├── scripts                          # Slurm launchers and offline evaluation scripts
├── third_party                      # OpenPI and RoboMeter submodules
├── weaver                           # Core WEAVER package and training/generation entrypoints
│   ├── datasets                     # Runtime DROID-style dataloaders
│   ├── utils                        # Config, checkpointing, evaluation, and metric utilities
│   └── wm                           # Latent flow world model, encoders, decoders, and transformer blocks
├── pyproject.toml                   # Package metadata and dependencies
└── README.md
```

## 💾 Datasets

WEAVER expects preprocessed DROID-style trajectories with actions, states, language features, rewards, normalization statistics, and either view videos or precomputed SD3 latents.
Reward-enriched annotations are loaded from `annotation_rewards/<split>/` by
default.

Preprocess a raw DROID download into the format expected by WEAVER:

```bash
python datasets/preprocess_droid.py \
  --data_root /path/to/raw_droid \
  --output_root /path/to/preprocessed_droid
```

See the [dataset preprocessing guide](datasets/README.md) for the expected
folder structure, parallel preprocessing, and normalization statistics.

To collect and preprocess your own custom robot trajectories, see:
- [Collecting custom trajectories](#collecting-custom-trajectories) — logging real rollouts with `panda_log.py`
- [Preprocessing custom OOD data](datasets/README.md#preprocessing-customized-ood-data) — converting raw logs into WEAVER format with SD3 latents and CLIP text features

## 🚀 Training WEAVER

By default, normalization statistics are loaded from
`<dataset.path>/norm_stats_relabel.json`. Released checkpoints should bundle
this file and set `dataset.norm_stats_path=/path/to/model/norm_stats_relabel.json`
when training or running inference.

Pretrain from scratch:

```bash
DATASET_PATH=/path/to/preprocessed_droid \
SCRATCH_DIR=/path/to/output/model_dir \
sbatch scripts/pretrain.sh
```

Finetune from a checkpoint:

```bash
PRETRAINED_DIR=/path/to/pretrained/logs/chkpts \
DATASET_PATH=/path/to/finetune_data \
EXP_NAME=weaver_finetune \
FINETUNE_SUFFIX=finetune \
sbatch scripts/finetune.sh
```

Run ReFlow post-training to distill a multi-step teacher into a faster student rollout:

```bash
PRETRAINED_DIR=/path/to/teacher/logs/chkpts \
PRETRAINED_CKPT_NAME=checkpoint.pt \
DATASET_PATH=/path/to/preprocessed_droid \
EXP_NAME=weaver_reflow \
FINETUNE_SUFFIX=reflow \
sbatch scripts/reflow.sh
```

All training launchers use four H100 GPUs with distributed data parallelism.

## 🔮 Inference

### Basic Inference

Generate rollout views and videos:

```bash
python -m weaver.generate_views \
  --checkpoint /path/to/logs/chkpts \
  --output-dir /path/to/eval_output \
  --split val \
  --use-real-history \
  --overrides \
    dataset.path=/path/to/eval_dataset \
    model.val_steps=27 \
    eval_horizon=5 \
    eval_bootstrap=5 \
    inference.pyramid_stagger_width=1 \
    inference.pyramid_schedule=cosine
```

### ⏱️ Inference Schedules

The main inference controls are:

- `model.val_steps`: number of flow denoising steps.
- `eval_horizon`: number of future frames generated per chunk.
- `eval_bootstrap`: number of generated frames fed back before the next chunk.
- `inference.pyramid_schedule`: denoising schedule used during rollout (`linear`, `cosine`, `power`, `sigmoid`).
- `inference.pyramid_stagger_width`: offset between frame-wise denoising schedules.

For staggered inference, the effective number of function evaluations is:

```text
NFE = val_steps + eval_horizon * pyramid_stagger_width
```

Typical settings:

```bash
# Lockstep, all future frames share the same noise level.
model.val_steps=8 eval_horizon=5 inference.pyramid_stagger_width=0

# Staggered pyramid rollout.
model.val_steps=45 eval_horizon=5 inference.pyramid_stagger_width=1
```

Slurm launchers for generation workflows are available in [`scripts/`](./scripts/).

## 📊 World Model Evaluation

Compute FID, FVD, and LPIPS from an evaluation output folder:

```bash
EVAL_DIR=/path/to/eval_output \
sbatch scripts/compute_eval_metrics.sh
```

The folder may be the evaluation root or its `views/` subdirectory. It must
contain the saved `gt_*.npy` and `pred_*.npy` camera views.
All trajectories with complete GT and prediction files are included by default.

---

## 🤖 Policy Evaluation
To evaluate policy success rates with the world model, we replay trajectories with WEAVER and save the generated rollouts. We manually label the generated data to get success rate with the world model. We also provide per-frame reward estimates, which can be used to evaluate and rank robot policies without real-world rollouts.

The following script encodes initial observation frames into WEAVER's latent space, then autoregressively predicts future frames conditioned on recorded actions, scores the rollout with the learned reward head, and saves a side-by-side GT-vs-prediction comparison video with the reward curve overlaid.

```bash
replay_options=(
  CHECKPOINT=/path/to/chkpts       # Directory containing checkpoint.pt
  DATASET_PATH=/path/to/dataset    # Preprocessed DROID-style dataset
  OUTPUT_DIR=/path/to/output       # Generated videos and reward files
  TRAJ_IDS="10 11 12"              # Set to "" to use NUM_TRAJS instead
  NUM_TRAJS=10                     # Used only when TRAJ_IDS is empty
  SAVE_REWARDS=1                   # Save rewards into annotation JSON files
  SHOW_REWARD=0                    # Overlay the reward curve on videos
)

env "${replay_options[@]}" bash scripts/replay_reward.sh
```

## 🎛️ Policy Finetuning

### 📹 Collecting Custom Trajectories

`third_party/openpi/examples/droid/panda_log.py` runs the policy on a real
Franka Panda robot and logs each rollout as a video + annotation JSON in the
same format expected by `preprocess_droid_ood.py`.

**Prerequisites:** start the OpenPI policy server on a GPU machine before running:

```bash
# On the remote GPU machine — run from the openpi repo root
uv run scripts/serve_policy.py --env DROID
```

**Log trajectories** (run from `third_party/openpi`):

```bash
uv run examples/droid/panda_log.py \
 --folder /path/to/your/task_folder \
 --external_camera left \
 --remote_host <server-ip>
```

Each rollout:
1. Prompts for a natural-language instruction
2. Executes the policy and records joint positions, gripper state, and all three camera views
3. Asks whether the rollout succeeded (`y` / `n`)
4. If confirmed, saves to `<folder>/videos/<id>.mp4` and `<folder>/annotations/<id>.json`
5. Optionally resets the robot and loops for the next trajectory

Key parameters:

| Parameter | Default | Description |
|-----------|---------|-------------|
| `--folder` | `logs/` | Output directory for videos and annotations |
| `--external_camera` | required | External camera fed to the policy: `left` or `right` |
| `--remote_host` | `0.0.0.0` | IP address of the OpenPI policy server |
| `--remote_port` | `8000` | Port of the policy server |
| `--max_timesteps` | `1000` | Maximum steps per rollout |
| `--open_loop_horizon` | `9` | Steps executed per predicted action chunk |
| `--interface_cfg` | `charmander.yml` | Deoxys robot interface config |

The saved `<folder>/` can be passed directly as an `--input_roots` entry to
`preprocess_droid_ood.py` — set `--tasks none` since there is no task sub-directory:

```bash
python -m datasets.preprocess_droid_ood \
    --input_roots /path/to/your/task_folder \
    --output_root /path/to/weaver_format_data \
    --data_type train \
    --tasks none
```

---

### ⚗️ Synthetic Data Generation
To finetune policies with WEAVER, we generate synthetic data with the base policy ($\pi_{0.5}$) and best-of-N sampling.
The segments are obtained by replaying the base policy within WEAVER for M chunks. Each segment is scored with the advantage computed using the latent reward and critic heads for the entire sequence. The best action candidate sequence is added to the synthetic dataset. The synthetic dataset can be combined with real data to finetune the base policy.

**Prerequisites: start the OpenPI server on a GPU machine (e.g. A6000) before running:**

```bash
# On the remote GPU machine — run from the openpi repo root
uv run scripts/serve_policy.py --env DROID --num-samples 5
```

Then set `PI_HOST` to that machine's IP/hostname when launching the script below.

```bash
synth_options=(
 CHECKPOINT=/path/to/chkpts # Directory containing checkpoint.pt
 DATASET_PATH=/path/to/dataset # Preprocessed source dataset
 OUTPUT_DIR=/path/to/output # Generated trajectories and videos
 DYNAMICS_MODEL=/path/to/model.pth # Ctrl-World dynamics checkpoint
 NUM_TRAJECTORIES=1000 # Number of trajectories to generate
 NUM_SAMPLES=5 # Best-of-N policy samples per chunk
 NUM_CHUNKS=4 # Imagination chunks per trajectory
 OPEN_LOOP_HORIZON=9 # Control steps executed per chunk
 SELECTION_CRITERION=advantage # Candidate ranking: reward or advantage
 FILTER_EPISODE_ID="stack" # Episode-ID substring; "" disables filtering
 PI_HOST=<server-ip> # OpenPI policy server hostname or IP
 PI_PORT=8000 # OpenPI policy server port
 DEBUG=0 # Save per-sample debug videos
)

env "${synth_options[@]}" bash scripts/synth_data_gen.sh
```

---

### 📦 Converting Synthetic Data for Finetuning

After generating synthetic data (or to package any mix of real + synthetic data for fine-tuning), use the three-step pipeline below.

#### Step 1 — Convert to DROID layout

`convert_synthetic_data_to_droid.py` reads one or more root directories and
assembles a unified DROID-format output folder.  Each `--tasks` entry is a
sub-folder name relative to the root (pass `none` to read `annotations/` and
`videos/` directly from the root, useful for real-data folders with no task
sub-directory).

```bash
# Synthetic chunks only
python third_party/openpi/examples/droid/convert_synthetic_data_to_droid.py \
    --input-dirs /path/to/synthetic_data_folder \
    --output-dir ../data/data_synthetic_droid \
    --tasks cup_task marker_task

# Mix: synthetic chunks + real data sub-folder
python third_party/openpi/examples/droid/convert_synthetic_data_to_droid.py \
    --input-dirs /path/to/synthetic_data_folder /path/to/real_data_folder \
    --output-dir ../data/data_mixed_droid_all \
    --tasks cup_task marker_task

# Real data only (annotations/ sits directly under the root)
python third_party/openpi/examples/droid/convert_synthetic_data_to_droid.py \
    --input-dirs /path/to/real_data_folder \
    --output-dir ../data/data_real_droid \
    --tasks none
```

Output layout (one folder per trajectory, plus a shared annotations file):

```
<output-dir>/
 traj_0000/
 trajectory.h5
 recordings/MP4/camera_{0,1,2}.mp4
 metadata_0.json
 traj_0001/
 ...
 aggregated-annotations-030724.json
```

#### Step 2 — Convert DROID layout to LeRobot format

`convert_synthetic_droid_data_to_lerobot.py` ingests the output of Step 1 and
writes a LeRobot dataset.  Pass `--repo_name` to set both the local dataset
name and the HuggingFace Hub repo ID.  Add `--push_to_hub` to upload directly.

```bash
python third_party/openpi/examples/droid/convert_synthetic_droid_data_to_lerobot.py \
    --data_dir ../data/data_mixed_droid_all \
    --repo_name your-hf-username/your-dataset-name
```

#### Step 3 — Fine-tune π₀.₅ on the converted dataset

> **Before running:** open `third_party/openpi/src/openpi/training/config.py` and
> update the `repo_id` on line 905 to match the `--repo_name` you used in Step 2:
>
> ```python
> # third_party/openpi/src/openpi/training/config.py  (line 905)
> repo_id = "your-hf-username/your-dataset-name",  # ← change this
> ```

Then launch fine-tuning from the `third_party/openpi` directory:

```bash
cd third_party/openpi
uv run scripts/train.py pi05_droid_finetune_real_syn_adv \
    --exp-name=droid-20k-real-syn-finetune \
    --overwrite
```

---

## 🕹️ Test-time Planning
WEAVER post-trained with rectified flow objective enables test-time steering with reduced inference cost and efficient generations in the latent space. For each observation, the script queries the base policy ($\pi_{0.5}$) for N action samples, rolls and scores them using advantage computed with WEAVER. The best action chunk is executed on the robot. This closed-loop steering improves real-world task success without any additional training.

**Prerequisites: start the OpenPI server on a GPU machine (e.g. A6000) before running:**

```bash
# On the remote GPU machine — run from the openpi repo root
uv run scripts/serve_policy.py --env DROID --num-samples 5
```

Then set `PI_HOST` to that machine's IP/hostname when launching the script below.

```bash
planning_options=(
 CHECKPOINT=/path/to/chkpts # Directory containing checkpoint.pt
 OUTPUT_DIR=/path/to/output # Output videos and debug grids
 DYNAMICS_MODEL=/path/to/model.pth # Ctrl-World dynamics checkpoint
 TASK="stack the red block" # Robot language instruction
 NUM_SAMPLES=8 # Best-of-N policy samples per chunk
 OPEN_LOOP_HORIZON=9 # Control steps before re-planning
 SELECTION_CRITERION=advantage # Candidate ranking: reward or advantage
 MAX_STEPS=700 # Maximum control steps per trial
 PI_HOST=<server-ip> # OpenPI policy server hostname or IP
 PI_PORT=8000 # OpenPI policy server port
)

env "${planning_options[@]}" bash scripts/steer_pi_policy.sh
```

---

## 📚 Citation

If you use WEAVER, please cite the corresponding paper once available.

```bibtex
@article{jain2026weaver,
  title={WEAVER: Efficient World Models for Robot Video Prediction},
  author={Arnav Kumar Jain and Yilin Wu and Jesse Farebrother and Gokul Swamy and Andrea Bajcsy},
  journal={CoRR},
  volume={abs/2606.13672},
  year={2026}
}
```

---

## 🙏 Acknowledgements

WEAVER is developed from the open-source video foundation model [Stable Video Diffusion](https://github.com/Stability-AI/generative-models) and multi-view world model [Ctrl-World](https://ctrl-world.github.io/). The VLA model used in this repository is from [openpi](https://github.com/Physical-Intelligence/openpi).

WEAVER

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README