HyperAI超神経

ホームニュース最新論文チュートリアルデータセット百科事典 SOTA LLMモデル GPU ランキング学会

サイトについて

日本語

HyperAI超神経

Autonomous Driving On Carla Leaderboard

評価指標

Driving Score

Infraction penalty

Route Completion

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	Driving Score	Infraction penalty	Route Completion	Paper Title	Repository
TransFuser	61.181	0.714	86.694	TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous Driving
TF++ WP	66.32	0.84	78.57	Hidden Biases of End-to-End Driving Models
TCP	75.14	0.87	85.63	Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
GRIAD	36.79	0.6	61.85	GRI: General Reinforced Imitation and its Application to Vision-Based Autonomous Driving	-
Transfuser	16.93	0.42	51.82	Multi-Modal Fusion Transformer for End-to-End Autonomous Driving
TCP (Reproduced)	47.91	0.77	65.73	Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
MaRLn	24.98	0.52	46.97	End-to-End Model-Free Reinforcement Learning for Urban Driving using Implicit Affordances
ReasonNet	79.95	0.89	89.89	ReasonNet: End-to-End Driving with Temporal and Global Reasoning	-
TransFuser (Reproduced)	55.04	0.63	89.65	TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous Driving
InterFuser	76.18	0.84	88.23	Safety-Enhanced Autonomous Driving Using Interpretable Sensor Fusion Transformer
Transfuser+	34.58	0.56	69.84	-	-
LBC	8.94	0.73	17.54	Learning by Cheating
Latent TransFuser	45.20	0.72	66.31	TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous Driving
World on Rails	31.37	0.56	57.65	Learning to drive from a world on rails
CILRS	5.37	0.55	14.40	Exploring the Limitations of Behavior Cloning for Autonomous Driving
InterFuser (Reproduced)	34.15	0.45	74.79	Safety-Enhanced Autonomous Driving Using Interpretable Sensor Fusion Transformer
NEAT	21.83	0.65	41.71	NEAT: Neural Attention Fields for End-to-End Autonomous Driving
Learning From All Vehicles (LAV)	61.846	0.640	94.459	Learning from All Vehicles

0 of 18 row(s) selected.