2024年12月11日 · In this paper, we propose a unified Generative Pre-training for Driving (GPD-1) model to accomplish all these tasks altogether without additional fine-tuning. We represent each scene with ego, agent, and map tokens and formulate autonomous driving as a unified token generation problem.

GPD-1：驾驶生成预训练. 驾驶生成预训练 (GPD-1) 模型采用两阶段训练过程，为自动驾驶模拟和规划任务构建坚实的基础。首先训练 Map VQ-VAE 潜token化器，采用 L1 误差作为地图线位置，采用二元交叉熵 (BCE) 评估地图直线可见性，如 SLEDGE [8] 中定义的那样。

2024年12月12日 · GPD-1 proposes a unified approach that seamlessly accomplishes multiple aspects of scene evolution, including scene simulation, traffic simulation, closed-loop simulation, map prediction, and motion planning, all without additional fine-tuning.

gpd-1：生成预训练驾驶. 作者的生成预训练驾驶（gpd-1）模型使用两阶段训练过程来构建自动驾驶模拟和规划任务的坚实基础。作者首先训练地图vq-vae潜在标记器，采用l1误差用于地图线位置和二元交叉熵（bce）来评估地图线可见性，如sledge中定义。

In this paper, we propose a unified Generative Pre-training for Driving (GPD) model to ac-complish all these tasks altogether without additional fine-tuning. We represent each scene with ego, agent, and map tokens and formulate autonomous driving as a unified to-ken generation problem. We adopt the autoregressive trans-

In this paper, we propose a unified Generative Pre-training for Driving (GPD) model to accomplish all these tasks altogether without additional finetuning. We represent each scene with ego, agent, and map tokens and formulate autonomous driving as a unified token generation problem.

2025年1月8日 · gpd-1：生成预训练驾驶. 作者的生成预训练驾驶（gpd-1）模型使用两阶段训练过程来构建自动驾驶模拟和规划任务的坚实基础。作者首先训练地图vq-vae潜在标记器，采用l1误差用于地图线位置和二元交叉熵（bce）来评估地图线可见性，如sledge中定义。

Our model adapts the GPT-like architecture for autonomous driving scenarios with two key innovations: 1) a 2D map scene tokenizer based on VQ-VAE that generates discrete, high-level representations of the 2D BEV map, and 2) a hierarchical quantization agent tokenizer to encode agent information.

作者提出了一个统一的生成预训练驾驶（GPD-1）模型来统一这些元素。他们将地图、智能体和自我车辆编码为一组统一的token标记，使其能够将场景演化表述为场景标记的生成预测。 原文链接：七大任务SOTA！北大&清华GPD-1：专为智驾设计的生成式预训练

The unique Oculink interface of GPD G1 provides an effective bandwidth of 63Gbps, allowing for discrete graphics expansion for GPD products including the WIN Max 2 2023, providing users with the best gaming and office experience!