You can deploy this model and start earning money today!

SberRoboticsCenter/
GreenVLA-2b-base

Name: GreenVLA-2b-base
Availability: InStock

vision-language-action

SberRoboticsCenter/Qwen3-VL-2B-Instruct-action

finetune

apache-2.0

Views

No views yet

GreenVLA-2b-base

Staged Vision-Language-Action Model for Generalist Robots

Sber Robotics Center · Manipulation Team

arXiv

Project Page

Code

Overview

GreenVLA-2b-base is the lightweight base checkpoint of the Green-VLA family — a ~2B-parameter Vision-Language-Action model pretrained on both general-domain and robotics data (3,000+ hours of demonstrations across multiple embodiments).

This checkpoint combines:

VLM capabilities — Visual Question Answering, object pointing, bounding box prediction, and scene description.
Autoregressive action prediction — FAST token-based action generation for discrete control.
Flow-matching action expert — A continuous action head for smooth, high-frequency trajectory generation.

Use this checkpoint when you need a smaller model footprint for fine-tuning or deployment on resource-constrained hardware. For best performance, consider GreenVLA-5b-base.

Architecture

Component	Details
VLM Backbone	Qwen3-VL-2B-Instruct (vision encoder + language model)
Action Expert	Flow-matching transformer operating in a reduced hidden space
Action Tokenizer	FAST tokenizer for autoregressive action prediction
Total Parameters	~2B

Training Curriculum

This checkpoint corresponds to the Base stage of the Green-VLA curriculum:

Stage	Name	Status
L0	Foundational VLM pretraining	✓
L1	Multimodal grounding (VQA, pointing, bbox)	✓
R0	Multi-embodiment robotics pretraining	✓
R1	Embodiment-specific adaptation	—
R2	RL policy alignment	—

Quick Start

Installation

bash

git clone https://github.com/greenvla/GreenVLA.git
cd GreenVLA
uv sync  # or: pip install -e .

Action Inference

python

import numpy as np
import torch
from lerobot.common.policies.factory import load_pretrained_policy
from lerobot.common.utils.torch_observation import (
    move_dict_to_batch_for_inference,
    torch_preprocess_dict_inference,
)

# 1. Load policy and transforms.
policy, input_transforms, output_transforms = load_pretrained_policy(
    "SberRoboticsCenter/GreenVLA-2b-base",
    data_config_name="bridge",
)
policy.to("cuda").eval()

# 2. Build an observation (replace with real sensor data).
raw_obs = {
    "observation/state": np.random.rand(8).astype(np.float32),  # x y z roll pitch yaw _pad_ gripper
    "observation/image": np.random.randint(0, 256, size=(224, 224, 3), dtype=np.uint8),
    "prompt": "pick up the green block and place it on the plate",
}

# 3. Transform, preprocess, and batch.
obs = input_transforms(raw_obs)
obs = torch_preprocess_dict_inference(obs)
batch = move_dict_to_batch_for_inference(obs, device="cuda")

# 4. Predict actions and post-process.
with torch.inference_mode():
    raw_actions = policy.select_action(batch).cpu().numpy()

actions = output_transforms(
    {"actions": raw_actions, "state": batch["state"].cpu().numpy()}
)["actions"]
# actions shape: (action_horizon, 7) — [x, y, z, roll, pitch, yaw, gripper]

See examples/example_inference_bridge.py for the full runnable script with argument parsing.

VLM Inference (VQA, Pointing, BBox)

The base model retains full VLM capabilities:

python

from PIL import Image
from lerobot.common.policies.factory import load_pretrained_policy

# Load without data transforms
policy, _, _ = load_pretrained_policy(
    "SberRoboticsCenter/GreenVLA-2b-base",
    data_config_name=None,
)
policy = policy.to("cuda").eval()

# Access the processor and model directly
processor = policy.model.processor
image = Image.open("scene.jpg")

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": image},
            {"type": "text", "text": "Describe what the robot should do next."},
        ],
    }
]

inputs = processor.apply_chat_template(
    messages, tokenize=True, add_generation_prompt=False,
    return_dict=True, return_tensors="pt",
    padding_side="left", padding="max_length", max_length=256,
    images_kwargs={"do_resize": True},
).to("cuda")

generated_ids = policy.model.model.generate(
    **inputs, max_new_tokens=256, do_sample=False, use_cache=False,
)

generated_ids_trimmed = [
    out[len(inp):] for inp, out in zip(inputs.input_ids, generated_ids)
]
print(processor.batch_decode(generated_ids_trimmed, skip_special_tokens=True)[0])

Model Family

Model	Stage	Params	Description	Link
GreenVLA-2b-base	Base	2B	Base pretrained (lightweight)	You are here
GreenVLA-5b-base	Base	5B	Base pretrained (recommended)	Hub
GreenVLA-5b-R1-bridge	R1	5B	Fine-tuned on Bridge (WidowX)	Hub
GreenVLA-5b-R2-bridge	R2	5B	RL-aligned on Bridge (WidowX)	Hub
GreenVLA-5b-R1-fractal	R1	5B	Fine-tuned on Fractal (Google Robot)	Hub

Citation

bibtex

@misc{apanasevich2026greenvlastagedvisionlanguageactionmodel,
    title   = {Green-VLA: Staged Vision-Language-Action Model for Generalist Robots},
    author  = {I. Apanasevich and M. Artemyev and R. Babakyan and P. Fedotova and
               D. Grankin and E. Kupryashin and A. Misailidi and D. Nerus and
               A. Nutalapati and G. Sidorov and I. Efremov and M. Gerasyov and
               D. Pikurov and Y. Senchenko and S. Davidenko and D. Kulikov and
               M. Sultankin and K. Askarbek and O. Shamanin and D. Statovoy and
               E. Zalyaev and I. Zorin and A. Letkin and E. Rusakov and
               A. Silchenko and V. Vorobyov and S. Sobolnikov and A. Postnikov},
    year    = {2026},
    eprint  = {2602.00919},
    archivePrefix = {arXiv},
    primaryClass  = {cs.RO},
    url     = {https://arxiv.org/abs/2602.00919},
}