Open techshoww opened 2 years ago
I believe the pre-trained model has been trained on KITTI-train split only(about 3K images), the benchmark result was conducted with the model trained on KITTI-trainval(7.5K images).
我已经在kitti测试数据集上运行了官方的MonoFlex预训练模型和代码,得到了以下分数:
以上分数与MonoFlex的官方分数之间存在很大差距。那么,我怎样才能在kitti测试铅板上重现分数呢?如何在kitti测试数据集上生成预测? 请问你在kitti测试集上有达到论文里面的效果吗?
I have run the official MonoFlex pretrained model and code on kitti test dataset , got below score:![image](https://user-images.githubusercontent.com/82649220/167620988-84816f57-03e1-45a0-8655-e271845b5e55.png)
There is a big gap between above score and MonoFlex's official score. So, how can i reproduce the score on kitti test leadboard ? How to generate predictions on kitti test dataset?