Help needed in Temp anomaly detection

Hi all, I’m trying to do an anomaly detection with temperature data. I have collected almost11hrs of data and uploaded to the EI. and here is my impulse setup.

But when I’m trying to do a live classification, it’s fails in my training data itself. I don’t know what i am missing.

Here is my EI Model. freezer-monitor - Dashboard - Edge Impulse

Hi @IoBytes,

The live classification takes data from your test set, not the training one.
Could you also explain what kind of anomalies you are looking for in your project?