diff --git a/README.md b/README.md index d229b3f..06f107e 100644 --- a/README.md +++ b/README.md @@ -191,6 +191,7 @@ We provide a leaderboard to present the driving performance evaluation of drivin - [x] Inference Code - [x] Training Code - [x] Pretrained Weights + - [ ] Code for nuPlan - [ ] Pretrained Model trained on nuScenes + nuPlan - [ ] Evaluation Code - [ ] Development Tutorial diff --git a/WorldDreamer/README.md b/WorldDreamer/README.md index 25fd409..5fb8d32 100644 --- a/WorldDreamer/README.md +++ b/WorldDreamer/README.md @@ -12,7 +12,7 @@ A Controllable Autoregressive Generative Network ## Changelog -[2024-09-15] `WorldDreamer` v1.0 is released. We now support single-frame autoregressive generator on `nuScenes` and `nuPlan` datasets. +[2024-09-15] `WorldDreamer` v1.0 is released. We now support single-frame autoregressive generator on `nuScenes` dataset. ## Installation diff --git a/WorldDreamer/docs/GETTING_STARTED.md b/WorldDreamer/docs/GETTING_STARTED.md index a5c2897..4c9acf0 100644 --- a/WorldDreamer/docs/GETTING_STARTED.md +++ b/WorldDreamer/docs/GETTING_STARTED.md @@ -10,7 +10,7 @@ The following codes are all run in the `WorldDreamer` folder unless otherwise sp ## Dataset Preparation (You can skip this step if you only want to use the simulation.) -Currently we provide the dataloader of [nuScenes dataset](#nuscenes-dataset) and [nuPlan dataset](#nuplan-dataset). +Currently we provide the dataloader of [nuScenes dataset](#nuscenes-dataset). The dataloader for [nuPlan dataset](#nuplan-dataset) will be released soon. ### nuScenes Dataset @@ -96,7 +96,7 @@ python -m tools.create_data nuscenes \ ``` > 🌻 You can download the `.pkl` files from [huggingface](https://huggingface.co/datasets/jokester-yxm/DriveArena_data). -### nuPlan Dataset +### nuPlan Dataset (Coming soon...) - To ensure a likely even distribution of the training data, we selected 64 logs from the NuPlan dataset. This selection includes 21 logs recorded in Las Vegas, 21 logs recorded in Pittsburgh, 11 logs recorded in Boston, and 11 logs recorded in Singapore. The names of the selected logs are listed under the `dreamer_train` and `dreamer_val` categories in [nuplan.yaml](../tools/data_converter/nuplan.yaml). Please download the official [nuPlan dataset](https://www.nuscenes.org/nuplan#download) and organized the files as follows: @@ -196,5 +196,6 @@ Test on the demo data, which is crop from the OpenStreetMap: python tools/test.py runner.validation_index=demo resume_from_checkpoint=path/to/your/weight ``` ## Todo +- [ ] release code for nuPlan - [ ] check tensorboard code - [x] check map visualization code \ No newline at end of file diff --git a/WorldDreamer/docs/GETTING_STARTED_CN.md b/WorldDreamer/docs/GETTING_STARTED_CN.md index 7c71696..2ddd5d0 100644 --- a/WorldDreamer/docs/GETTING_STARTED_CN.md +++ b/WorldDreamer/docs/GETTING_STARTED_CN.md @@ -10,7 +10,7 @@ ## 数据集准备 (使用仿真器时可以跳过这一步骤) -目前我们提供了 [nuScenes 数据集](#nuscenes-数据集) 和 [nuPlan 数据集](#nuplan-数据集) 的数据加载器。 +目前我们提供了 [nuScenes 数据集](#nuscenes-数据集)的dataloader。[nuPlan dataset](#nuplan-dataset)的dataloader即将发布。 ### nuScenes 数据集 @@ -96,7 +96,7 @@ python -m tools.create_data nuscenes \ └── val_200x200_12Hz_interp.h5 ``` > 🌻 你可以从 [huggingface](https://huggingface.co/datasets/jokester-yxm/DriveArena_data) 下载到`.pkl`文件。 -### nuPlan 数据集 +### nuPlan 数据集 (Coming soon) - 为确保训练数据的均匀分布,我们从 NuPlan 数据集中选择了 64 个log。此选择包括在拉斯维加斯记录的 21 个日志、在匹兹堡记录的 21 个日志、在波士顿记录的 11 个日志以及在新加坡记录的 11 个日志。这些日志的名称列在 [nuplan.yaml](../tools/data_converter/nuplan.yaml) 中的 `dreamer_train` 和 `dreamer_val` 分类下。请下载官方的 [nuPlan 数据集](https://www.nuscenes.org/nuplan#download) 并将文件组织如下: ``` @@ -199,6 +199,6 @@ python tools/test.py runner.validation_index=demo resume_from_checkpoint=path/to ``` ## Todo - +- [ ] nuPlan数据集的代码 - [ ] 检查 tensorboard 代码 - [x] 检查地图可视化代码