Project Overview provides a bird's-eye view of the progress of your machine learning projects. On the project overview dashboard, you can manage and track key information about your project:
- Key Metrics: Keep track of essential evaluation metrics of your experiments such as accuracy, loss, and MSE.
- Sample Media: Log images, audio, and other media from your experiment and explore your model's prediction results to compare your experiments visually.
- Starred Experiments: Star and keep track of meaningful experiments.
- Project Notes: Make a note of important information about your project and share it with your teammates – similar to README.md of Git codebase.
VESSL AI automatically marks metrics of best-performing experiments as key metrics. You can also manually bookmark key metrics and keep track of your model's meaningful evaluation metrics.
To add or remove Key Metrics
- 1.Click the settings icon() on top of the Key Metrics card.
2. Select up to 4 metrics and choose whether your goal is to minimize or maximize the target value.
- If you select Minimize, an experiment with the smallest target value will be updated to the key metric charts.
- If you select Maximize, an experiment with the greatest target value will be updated to the key metric chart.
You can log images or audio clips generated from your experiment to explore your model's prediction results and make visual (or auditory) comparisons.
To see the media file, select an experiment and specify the media type using the dropdown menu on the upper right corner of Sample Media card.
You can mark important experiments as Starred Experiments to keep track of meaningful achievements in the project. Starred Experiments displayed with the tags and key metrics.
To star or unstar experiments
- 1.Go to the experiment tracking dashboard
- 2.Select experiments
- 3.Click 'Star' or 'Unstar' on the dropdown menu.
You can also star or unstar experiments on the experiment summary page.
Project Notes is a place for noting and sharing important information about the project together with your team. It works like README.md of your Git codebase.
To modify project note, click the settings icon(
) on top of Project Notes card. You will be given a markdown editor to update your notes.