Feature: Dummy Model

[DiogenesAnalytics]

Problem

When trying to determine the performance of a recently trained autoencoder it can be helpful to have a "baseline" to compare it to. This would require another autoencoder that is just randomly "reconstructing" the output from the inputs (i.e. it is not being trained to "learn" anything about how to compress the data, it is just randomly producing output based on input).

Solution

A dummy model is really what is needed for this specific problem.

References

• https://chat.openai.com/share/a965e1d0-324e-42f3-8068-0b6e758e18ed
• https://keras.io/2.15/api/layers/core_layers/lambda/

[DiogenesAnalytics]

Implemented with: 93b9452ab1f6601a0bb5020005945062d5cedd50