After creating a dataset and trying to create a model, the same error is constantly displayed. Although keypoints were created for all classes, all photos used in the dataset were annotated. Previously, I thought that the reason was that the classes were named in ukrainian letters, but after changing them to english transcription, the photos in the dataset began to display classes and keypoints correctly, but the error when creating the model did not disappear. Another strange thing is that fast training works and creates a model based on the dataset. Here is a link to the project, sign languages 2 Keypoint Detection Dataset and Pre-Trained Model by sign languages maybe someone has encountered this problem before
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.