# Suggestion mode If you accelerate the video annotator and confirm 20+ labels the AI model will turn on and start generating predictions and suggestions. You will see that the mode switches to suggestions, the little squares will turn on to indicate which part of your data the suggestions are pointing to and if you navigate to a suggestion you will see predictions. \ \ Predictions are the automatic bounding boxes that have a confidence bar under the label. The model keeps learning as you provide more labels and predictions will keep getting better over time. Suggestions are the thumbnails with crosses in the right corner, these are the frames that the AI is most confused about, labelling these frames will increase the performance of the object localizer faster than other frames. You can click the cross to dismiss a suggestion that you don't want to label. It takes the model some time to generate good suggestions as it has to first scan through the video dataset and score the frames. The suggestions you get in the first few minutes will not be as good as suggestions you get once the AI has had more time to scan through the data. Speed of scanning depends on your plan but lies between 10-50FPS. If you de-accelerate a project the model will turn off and you will lose the pre-computed suggestions. However the model training will not be impacted as the model automatically makes checkpoints. It will take some time for suggestions to get to a high quality after a reset of the AI. ![](/files/cAomo0d05hvA0lSUbafX) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.lodestar.ai/dataset-management/video-annotator/modes-while-annotating/suggestion-mode.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.