ig248 / kerashistoryplot

Plots training losses and metrics in Jupyter while training your `keras` model.

History plot for Keras model training

Plots training losses and metrics in Jupyter while training your keras model.

This project started as an attempt to extend livelossplot, but I realized that a rewrite was easier. Key improvements are:

• keeps full history, including losses for each batch in each epoch
• plots scatter of batch losses
• plots arbitrary log entries, such a lr (learning rate)
• simplifies codebase extendibility

Basic Usage

Install package from source:

pip install git+https://github.com/ig248/kerashistoryplot

Or from PyPi:

pip install kerashistoryplot

Import function:

from livehistoryplot.callbacks import PlotHistory

Use in callbacks:

model.fit(
    X, y,
    epochs=20,
    callbacks=[
        ReduceLROnPlateau(),
        PlotHistory(batches=True, n_cols=3, figsize=(15, 7))
    ]
)

More Examples

See notebooks/

Tests

Run

make dev-install
make test

References

Original livelossplot: https://github.com/stared/livelossplot

kerashistoryplot's Issues

Investigate `pyqtplot`