How to solve the `too many cavs reinitialize` problem? Just delete it?

[Chinese-Coding]

too many cavs reinitialize

Chinese

当我第一次按照您提供的README中执行本程序时, 在控制台出现了这样一行输出 too many cavs reinitialize, 并伴随着一些 errors. 我第一次时直接将那些不符合条件的时间戳的文件夹给移走了, 例如: dataset/OPV2V/train/2021_09_09_22_21_11 这个文件夹, 它里面总共有 6 辆车 (Figure 1). 当我今天下载好 OPV2V-H 数据集时, 我发现 OPV2V-H 数据集中保留了这些文件, 并做了相应的处理 (顺便提一嘴, 我发现了 OPV2V-H 有国内的镜像, 就是下载速度有些慢, 平均 3M/s, 我没有测试过直接下载的情况, 自我感觉良好). 所以, 总感觉有些不妥当, 想问一下大家是怎么解决的.

为了提高代码的鲁棒性, 我还在输出 too many cavs reinitialize 的地方添加了如下的代码 (Figure 2)

English

When I first executed this program according to the README you provided, a line of output too many cavs reinitialize appeared in the console, accompanied by some errors. The first time I directly moved the folders with timestamps that did not meet the conditions, for example: the folder dataset/OPV2V/train/2021_09_09_22_21_11, which has a total of 6 vehicles (Figure 1).

When I downloaded the OPV2V-H dataset today, I found that these files were retained in the OPV2V-H dataset and processed them accordingly (by the way, I found that OPV2V-H has a domestic mirror, but the download speed is a bit slow, averaging 3M/s. I haven't tested the direct download situation, but I feel good about it). So, I always feel that something is wrong, and I want to ask how everyone solves it.

In order to improve the robustness of the code, I also added the following code where too many cavs reinitialize is output (Figure 2)

Figure

Figure 1: There 6 cavs in `dataset/OPV2V/train/2021_09_09_22_21_11`

Figure 2

warnings.warn(f'In {scenario_folder}, there are too many cavs reinitialize, '
                                  f'Please check the data set to avoid errors')

[yifanlu0227]

Hi, 这个没有关系。

在yaml里的train_params.max_cav可以控制场景里最多的agent数量。这个通常被设置成5，如果数据集里一个scene 的agent数量大于5，它就会break，控制读取的agent数量为5。你可以忽略这个warning。