数据集问题

[yumu-173]

Prerequisite

• [X] I have searched Issues and Discussions but cannot get the expected help.
• [X] I have read the FAQ documentation but cannot get the expected help.
• [X] The bug has not been fixed in the latest version (master) or latest version (3.x).

Task

I'm using the official example scripts/configs for the officially supported tasks/models/datasets.

Branch

master branch https://github.com/open-mmlab/mmdetection

Environment

torch1.10.0 mmcv-full 1.3.17

Reproduces the problem - code sample

if ann.get('iscrowd', False): gt_bboxes_ignore.append(bbox) else: gt_bboxes.append(bbox) gt_labels.append(self.cat2label[ann['category_id']]) gt_masks_ann.append(ann['segmentation'])

Reproduces the problem - command or script

bash tools/dist_train.sh

Reproduces the problem - error message

当我运行dist_train.sh脚本时，需要rawData中的*_wo_ignore.json并且在读取注释json的过程中还需要注释的segmentation，这些是SODA数据集中没有提供的，并且因为训练过程比较长，想请问是否可以提供预训练的脚本，可以直接进行test

Additional information

1、更全面的注释文件 2、预训练模型

[shaunyuan22]

1. train_wo_ignore.json can be obtained by filtering the ignore annotations in train.json, likewise for val_wo_ignore.json and test_wo_ignore.json . We've updated the code and please refer to here.
2. segmentation will NOT be used in training detection models, namely you don't have to load gt_masks during loading SODA-D data, please check your training config.
3. We provide the specific configs for benckmarking models in our paper please refer to SODA-mmdet and SODA-mmrot. Releasing the pretrained models is not in our near-term plans and your understanding is much appreciated.