Live loss plotting during training¶

[1]:

if 'google.colab' in str(get_ipython()):
    !pip install git+https://github.com/ourownstory/neural_prophet.git['live'] # may take a while
    #!pip install neuralprophet # much faster, but may not have the latest upgrades/bugfixes

import pandas as pd
from neuralprophet import NeuralProphet

[2]:

data_location = "https://raw.githubusercontent.com/ourownstory/neuralprophet-data/main/datasets/"
df = pd.read_csv(data_location + "retail_sales.csv")

[3]:

df_train, df_val = NeuralProphet().split_df(df, valid_p=0.2)

INFO - (NP.df_utils._infer_frequency) - Major frequency MS corresponds to 91.126% of the data.
INFO - (NP.df_utils._infer_frequency) - Dataframe freq automatically defined as MS

[4]:

m = NeuralProphet()
metrics = m.fit(df_train, validation_df=df_val, progress="plot")

_images/Live_plot_during_training_4_0.png

log-SmoothL1Loss
        training                 (min:   -7.234, max:   -1.345, cur:   -7.234)
        validation               (min:   -4.842, max:   -1.321, cur:   -3.239)

[5]:

m = NeuralProphet()
metrics = m.fit(df_train, validation_df=df_val, progress="plot-all")

_images/Live_plot_during_training_5_0.png

MAE
        training                 (min: 6564.215, max: 253641.383, cur: 6577.867)
        validation               (min: 31846.620, max: 267939.478, cur: 63734.493)
RMSE
        training                 (min: 8668.035, max: 283905.474, cur: 8856.043)
        validation               (min: 34741.931, max: 300150.946, cur: 67346.074)
RegLoss
        RegLoss                  (min:    0.000, max:    0.000, cur:    0.000)
log-SmoothL1Loss
        training                 (min:   -7.233, max:   -0.494, cur:   -7.233)
        validation               (min:   -4.564, max:   -0.408, cur:   -3.241)

[ ]: