search

Vegeta2020 / SE-SSD

SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud, CVPR 2021.
Apache License 2.0
811 stars 128 forks source link

Error in loading checkpoint while training #82

Closed bhaskar-anand-iith closed 2 years ago

bhaskar-anand-iith commented 2 years ago

When I am trying to train the model by

python tools/train.py

I am getting following error

OSError: /mnt/proj50/zhengwu/saved_model/KITTI/proj52/megvii/second/pre_trained_model_2/epoch_60.pth is not a checkpoint file

Which is obvious as I am not giving path to the pretrained model epoch_60.pth. But I want to know why do we need a pretrained model if we want to train from scratch? If it is required, from where do we get epoch_60.pth?

Vegeta2020 commented 2 years ago

@bhaskar-anand-iith You may download the pre-trained CIA-SSD (see the link in its repository) as this checkpoint, and you also need to revise the checkpoint path in the config file.

bhaskar-anand-iith commented 2 years ago

Ok, thank you..

lacie-life commented 11 months ago

I downloaded the .pth file from this but met the new error:

"raise IOError("{} is not a checkpoint file".format(filename)) OSError: ~/Dataset/Kitti/saved_model/cia-ssd-model.pth is not a checkpoint file"

Can you help me?

Thanks!