Numeric Data Teachable Machine

1. Choose clue columns 2. Add the Label/Class 3. Train and predict

Set Up Your Dataset

Enter numeric clue columns first. The tool will always add one final column called Label/Class.

Dataset name

Clue columns

Type one numeric clue column per line. Numeric clues can be counts, measurements, or ratings like 0, 0.5, and 1.

Labels / Classes

Type one possible answer per line.

Initial training data

Optional. Type one example per line, separated by commas. Clues must be numbers. The last item is the Label/Class.

1. Add examples 2. Train TensorFlow 3. Predict

Training Data

Rows are examples. Clue columns are features. The last column is the Label/Class.

Add training example

Add training examples from CSV

Optional. Type one example per line, separated by commas. Clues must be numbers. The last item is the Label/Class.

What the model learned

Train the model to see patterns.

Prediction

Enter any number clues you know. Decimals are allowed, so values between training examples can still be predicted. Blank clues are treated as unknown.

Batch predictions

Optional. Type one row per line, separated by commas. Leave a clue blank when it is unknown.

Teacher language

Feature A number clue column, like Age, Temperature, or a 0-to-1 interest rating.

Label/Class The answer AI is learning.

Prediction The model's guess using the clue columns that are filled in.