In [1]:
import pandas as pd
import numpy as np
%matplotlib inline
In [2]:
initial = np.load("lstm_basic/initial.npy")
initial_df = pd.DataFrame(initial, columns=["initial"])
In [3]:
output1 = np.load("lstm_learning0.02/output.npy")
output1_df = pd.DataFrame(output1, columns=["output (learning: 0.02)"], index=range(len(initial), len(initial) + len(output1)))
losses1_df = pd.DataFrame(np.load("lstm_learning0.02/losses.npy"), columns=["epoch", "loss (learning: 0.02)"])
In [4]:
output2 = np.load("lstm_basic/output.npy")
output2_df = pd.DataFrame(output2, columns=["output (learning: 0.1)"], index=range(len(initial), len(initial) + len(output2)))
losses2_df = pd.DataFrame(np.load("lstm_basic/losses.npy"), columns=["epoch", "loss (learning: 0.1)"])
In [5]:
output3 = np.load("lstm_learning0.5/output.npy")
output3_df = pd.DataFrame(output3, columns=["output (learning: 0.5)"], index=range(len(initial), len(initial) + len(output3)))
losses3_df = pd.DataFrame(np.load("lstm_learning0.5/losses.npy"), columns=["epoch", "loss (learning: 0.5)"])
In [6]:
train = np.load("train_data/normal.npy")
train_df = pd.DataFrame(train[:, 0], columns=["train"])
In [7]:
merged = pd.concat([train_df, initial_df, output1_df, output2_df, output3_df])
merged.plot(figsize=(15, 5), grid=True, style=["-", "-", "-", "-", "k--"])
Out[7]:
In [8]:
merged = pd.merge(losses1_df, losses2_df)
merged = pd.merge(merged, losses3_df)
merged.plot(figsize=(15, 5), grid=True, logy=True, x="epoch")
Out[8]:
In [ ]: