Shilong Liu

这里是我们没有清理的失败的参数，您可以忽略这项，对模型没有实际影响。

Thanks for your question. We have fixed this bug in #28 . More details are available at https://github.com/IDEA-Research/GroundingDINO/issues/14#issuecomment-1491176248

这个看起来是环境而非我们项目本身的问题，您可以上网搜一下

1. we use avg-pooling in all the models. 2. All of our experiments leveraged the simplified ASL loss.

Thanks. The order will affect results as it views each input as a sentence. You can have a try: `dress . skirt . socks . shoes` The `.` with make...

Thanks to @xjtupanda for providing the solution!

Thanks for your suggestions. We will update batch inputs later. For now, you can modify the inputs in demo codes to enable batch inputs.

Do you evaluate the COCO pre-trained models on your custom dataset? You need to train a new model for your dataset.

As I do not know the custom dataset you used. I cannot reproduce the error. However, the error is raised by cocoapi. There are some similar problems in other projects,...

We did not meet the problem in our other custom datasets. You seem to have other costume codes, such as the `coco_detr_graffiti.py` file, which blocks us from reproducing it. I...