Get class-wise precision & recall in the Evaluation

[Nuwan1654]

Before Asking

• [X] I have read the README carefully. 我已经仔细阅读了README上的操作指引。

• [X] I want to train my custom dataset, and I have read the tutorials for finetune on your data carefully and organize my dataset correctly; 我想训练自定义数据集，我已经仔细阅读了训练自定义数据的教程，以及按照正确的目录结构存放数据集。

• [X] I have pulled the latest code of main branch to run again and the problem still existed. 我已经拉取了主分支上最新的代码，重新运行之后，问题仍不能解决。

Search before asking

• [X] I have searched the DAMO-YOLO issues and found no similar questions.

Question

Hi, I have trained the DAMO Yolo on my custom dataset with 10 classes, I need to get the class-wise precision and recall, how can I get that in the evaluation.

Additional

No response

[jyqi]

Thanks for your attention to DAMO-YOLO. You can get class-wise precision & recall in the evaluation by following steps:

Step1. Edit PythonAPI/pycocotools/coco.py and add following code in green area.

Step2. Edit PythonAPI/pycocotools/cocoeval.py and add following code in green area.

[Nuwan1654]

@jyqi Thanks for these instructions, I have another question though, I started the training using the following command python -m torch.distributed.launch --nproc_per_node=1 tools/train.py -f configs/damoyolo_tinynasL35_M.py, then after finishing it I got 20 checkpoints, saved, how can I select the best one,, like in yolov5/7, Is there any configuration to make it save only the best wights,

[jyqi]

Currently we do not have the feature of selecting the best checkpoint. You can find the best one by querying training logs.