Visualizations

mlop supports creating different visualizations depending on how you want to view your data, currently we support

• Metrics: These are simple line graphs of numerical data over time, e.g training loss over time
• Images: You can log images across time, this also includes support for logging matplotlib graphs over time. Potential use cases can be logging how well a image generative model improve its generation as the training progresses
• Audio: You can log audio across time, observing how the generated audio progresses over time
• Video: You can log video across time, observing how the generated video progresses over time
• Histograms: You can log histograms of your data, useful for understanding the distribution of your data

Additionally, we provide

• Model Graph: You can check the data flow and structure of your model, super useful for understanding the model architecture