concerto-large-lp.scannet20.pointcept

Model card for concerto-large-lp.scannet20.pointcept

A Concerto point cloud semantic segmentation model (joint 2D-3D representation encoder). Trained on ScanNet (20 classes).

Non-commercial. These weights are released by Pointcept/Concerto under CC BY-NC 4.0 and may be used for research and evaluation only.

Model Details

• Model Type: Point cloud semantic segmentation
• Model Stats:
  • Params (M): 207.7
  • Input channels: 9
  • Classes: 20
  • Features: 1728
• Dataset: ScanNet (20 classes)
• Metrics: mIoU 78.59, OA 92.29
• Paper: Concerto: Joint 2D-3D Self-Supervised Learning Emerges Spatial Representations
• Converted from: Pointcept/Concerto (CC-BY-NC-4.0)
• Library: torch-pointcloud

Install

pip install torch-pointcloud

This checkpoint also needs spconv and flash-attn, which need a build matching your torch and CUDA: see the installation guide.

Usage

import torch
import torch_pointcloud as tp
from torch_pointcloud.utils.data import collate

model, info = tp.create_model(
    "concerto-large-lp.scannet20.pointcept",
    pretrained=True,
    return_info=True,
)
model = model.cuda().eval()  # GPU-only kernels

# synthetic sample with the keys a dataset provides
num_points = 8192
sample = {
    "pos": torch.randn(num_points, 3),
    "color": torch.rand(num_points, 3) * 255,
    "normal": torch.randn(num_points, 3),
    "segment": torch.zeros(num_points, dtype=torch.long),
    "instance": torch.zeros(num_points, dtype=torch.long),
}
data = info["transform"](sample)
data = collate([data])
data = {key: value.cuda() for key, value in data.items()}

with torch.no_grad():
    logits = model(data.get("x"), data["pos_grid"], data["batch"])

Feature extraction

with torch.no_grad():
    features = model.forward_features(data.get("x"), data["pos_grid"], data["batch"])

model.reset_classifier(num_classes=0)
with torch.no_grad():
    features = model(data.get("x"), data["pos_grid"], data["batch"])  # (N, 1728)

Citation

@article{concerto2025,
  title   = {Concerto: Joint 2D-3D Self-Supervised Learning Emerges Spatial Representations},
  author  = {Yujia Zhang and Xiaoyang Wu and Yixing Lao and Chengyao Wang and Zhuotao Tian and Naiyan Wang and Hengshuang Zhao},
  journal = {arXiv preprint arXiv:2510.23607},
  year    = {2025}
}

@inproceedings{dai2017scannet,
  title   = {ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes},
  author  = {Angela Dai and Angel X. Chang and Manolis Savva and Maciej Halber and Thomas Funkhouser and Matthias Nießner},
  booktitle = {CVPR},
  year    = {2017}
}

@software{dujardin2026pytorchpointcloud,
  author  = {Arthur Dujardin},
  title   = {PyTorch PointCloud},
  year    = {2026},
  doi     = {10.5281/zenodo.22159632},
  url     = {https://github.com/arthurdjn/pytorch-pointcloud},
}