Hi, this is an excellent job. I have converted Argoverse1.1 to the nuScenes format so that my model can be trained and evaluated on it, as well as to validate new evaluation metrics. However, I currently don’t have enough time and computational resources to explore an appropriate number of training epochs and training strategies, such as the choice between one cycle, AdamW, or SGD. And I can’t find training strategies for detection models on the Argoverse1 dataset. Regarding leaderboard models, I also can’t find the related open-source code.
Hi, this is an excellent job. I have converted Argoverse1.1 to the nuScenes format so that my model can be trained and evaluated on it, as well as to validate new evaluation metrics. However, I currently don’t have enough time and computational resources to explore an appropriate number of training epochs and training strategies, such as the choice between one cycle, AdamW, or SGD. And I can’t find training strategies for detection models on the Argoverse1 dataset. Regarding leaderboard models, I also can’t find the related open-source code.