ModelPrésentéExterne

Topological Traps

TAU Algorithmic Robotics - Fall 2025/2026 - Daniel Simanovsky

Sourcehuggingface.co/DanielDDDS/topological-traps/tree/286e0cef0e52b414d58a0b101cc1fa111b384fe1

TypeOtherVersionRevision 286e0ceCClasse

Source du modèle

Type: Other
Taille du fichier: 280 MB
Version: vRevision 286e0ce
Mis à jour: 16 sept. 2026
Téléchargements: 0
Source: Hugging Face

Description de la source

Neural Prediction of Heading-Dependent Topological Traps

TAU Algorithmic Robotics - Fall 2025/2026 - Daniel Simanovsky

A U-Net model that predicts directional viability maps for non-holonomic robot navigation: given a 2D occupancy grid and robot dimensions, the model outputs four binary masks (N/S/E/W) indicating which pixels are topological traps - cells the robot can enter but cannot escape by translating in that heading without first rotating.

Models in this repo

File	Description	Val IoU
`models/viability_cardinal_best_iou.pth`	Cardinal model - 3-ch input, 4-ch binary output (N/S/E/W)	0.9793
`models/viability_continuous_angle_best_iou.pth`	Angle model - 5-ch input (sin/cos theta), 1-ch arbitrary heading	0.984
`configs/training_config.yaml`	Training configuration	-

Architecture

Backbone: U-Net with ResNet34 encoder (ImageNet pre-trained), ~24.5M parameters
Input: 3-channel tensor - binary occupancy grid + normalised robot L/512, W/512
Output: 4-channel sigmoid map, one per cardinal direction
Training: AdamW (lr=1e-4, wd=1e-4), CosineAnnealingLR, BCE+Dice loss, FP16, batch 16, 50 epochs

Results

Metric	Value
IoU (seen sizes)	0.978
IoU (unseen size 25x18)	0.953
Generalisation gap	0.025
Inference time	8.8 ms/map (GPU)
Oracle speedup	21.5x
PRM trap reduction	81%

Quick start

import torch, numpy as np
from huggingface_hub import hf_hub_download
import segmentation_models_pytorch as smp

ckpt_path = hf_hub_download(
    repo_id="DanielDDDs/topological-traps",
    filename="models/viability_cardinal_best_iou.pth"
)
model = smp.Unet(encoder_name="resnet34", encoder_weights=None, in_channels=3, classes=4)
state = torch.load(ckpt_path, map_location="cpu")
model.load_state_dict(state["model_state_dict"])
model.eval()

def predict(occupancy, robot_L=30, robot_W=20):
    x = np.zeros((1, 3, 512, 512), dtype=np.float32)
    x[0, 0] = occupancy
    x[0, 1] = robot_L / 512.0
    x[0, 2] = robot_W / 512.0
    with torch.no_grad():
        out = torch.sigmoid(model(torch.from_numpy(x)))
    return out[0].numpy()  # (4, 512, 512) N/S/E/W viability

viability = predict(occupancy_grid)
north_viable = viability[0] > 0.5

Dataset

Full processed maps and Oracle labels: DanielDDDs/topological-traps-dataset

Code

github.com/danielsddd/topological-traps-project

Compute

TAU CS SLURM cluster:

NVIDIA TITAN Xp (12 GB)
NVIDIA GeForce RTX 2080 (8 GB)

If you use this model in your research, please cite it as follows:

@misc{simanovsky2026neural,
  author = {Daniel Simanovsky},
  title  = {Neural Prediction of Heading-Dependent Topological Traps},
  year   = {2026},
  school = {Tel Aviv University},
}

Sources

1 source

Hugging FacePrincipalVersion 286e0cef0e52b414d58a0b101cc1fa111b384fe1

huggingface.co/DanielDDDS/topological-traps/tree/286e0cef0e52b414d58a0b101cc1fa111b384fe1

Vérifié 1 sept.

Artefacts du modèle

2 artefacts

models/viability_cardinal_best_iou.pth

pth · 280 MB · SHA-256 07e4b2304459…650d · Hugging Face

Télécharger

models/viability_continuous_angle_best_iou.pth

pth · 280 MB · SHA-256 1c8b72047306…1e75 · Hugging Face

Télécharger

Extraits de sources

2 extraits

Readme1 sept.huggingface.co/DanielDDDS/topological-traps

--- license: mit language: - en tags: - robotics - motion-planning - segmentation - unet - topological-traps - viability-prediction - non-holonomic pipeline_tag: image-segmentation datasets: - DanielDDDs/topological-traps-dataset --- # Neural Prediction of Heading-Dependent Topological Traps **TAU Algorithmic Robotics - Fall 2025/2026 - Daniel Simanovsky** A U-Net model that predicts directional viability maps for non-holonomic robot navigation: given a 2D occupancy grid and robot dimensions, the model outputs four binary masks (N/S/E/W) indicating which pixels are topological traps - cells the robot can enter but cannot escape by translating in that heading without first rotating. ## Models in this repo | File | Description | Val IoU | |---|---|---| | `models/viability_cardinal_best_iou.pth` | Cardinal model - 3-ch input, 4-ch binary output (N/S/E/W) | 0.9793 | | `models/viability_continuous_angle_best_iou.pth` | Angle model - 5-ch input (sin/cos theta), 1-ch arbitrary heading | 0.984 | | `configs/training_config.yaml` | Training configuration | - | ## Architecture - Backbone: U-Net with ResNet34 encoder (ImageNet pre-trained), ~24.5M parameters - Input: 3-channel tensor - binary occupancy grid + normalised robot L/512, W/512 - Output: 4-channel sigmoid map, one per cardinal direction - Training: AdamW (lr=1e-4, wd=1e-4), CosineAnnealingLR, BCE+Dice loss, FP16, batch 16, 50 epochs ## Results | Metric | Value | |---|---| | IoU (seen sizes) | 0.978 | | IoU (unseen size 25x18) | 0.953 | | Generalisation gap | 0.025 | | Inference time | 8.8 ms/map (GPU) | | Oracle speedup | 21.5x | | PRM trap reduction | 81% | ## Quick start ```python import torch, numpy as np from huggingface_hub import hf_hub_download import segmentation_models_pytorch as smp ckpt_path = hf_hub_download( repo_id="DanielDDDs/topological-traps", filename="models/viability_cardinal_best_iou.pth" ) model = smp.Unet(encoder_name="resnet34", encoder_weights=None, in_channels=3, classes=4) state = torch.load(ckpt_path, map_location="cpu") model.load_state_dict(state["model_state_dict"]) model.eval() def predict(occupancy, robot_L=30, robot_W=20): x = np.zeros((1, 3, 512, 512), dtype=np.float32) x[0, 0] = occupancy x[0, 1] = robot_L / 512.0 x[0, 2] = robot_W / 512.0 with torch.no_grad(): out = torch.sigmoid(model(torch.from_numpy(x))) return out[0].numpy() # (4, 512, 512) N/S...

Note1 sept.huggingface.co/DanielDDDS/topological-traps

Source context: 0 downloads · 0 likes · Pipeline image-segmentation · Repo DanielDDDS/topological-traps