Open Suncheng2022 opened 1 year ago
@Suncheng2022 The GLIP algorithm inherently has phrase groundingfunction. Currently mmdet only supports inference, fine-tuning will be supported in the future.
@Suncheng2022 The GLIP algorithm inherently has phrase groundingfunction. Currently mmdet only supports inference, fine-tuning will be supported in the future.
Can I use GLIP directly to generate a rougher object detection dataset using images with text descriptions?
May I ask whether mmdetection's GLIP will implement the "phrase grounding" function in their paper, the function is very attractive, it can automatically label a large amount of image data for object detection. I really hope you can add this feature! Best wishes!