#🔒 keras is returning nan

11 messages · Page 1 of 1 (latest)

narrow bluffBOT
#

@supple laurel

Python help channel opened

Remember to:

  • Ask your Python question, not if you can ask or if there's an expert who can help.
  • Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
  • Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

Closes after a period of inactivity, or when you send !close.
supple laurel
#

got deleted

#

code:```py
import os

os.environ["TF_ENABLE_ONEDNN_OPTS"] = "0"

import keras
import pandas as pd
from keras import layers
from sklearn.model_selection import train_test_split

from load_data import *

repro data

data = load_to_file(

"data/training_labels.RData", "data/full-parsed-data.shp", "train_labels"

)

data = gp.read_file("data/full-parsed-data.shp")
x = data[
[
"area",
"perimeter",
"con_hull",
"reock",
"len_width",
"polsby_pop",
]
]
y = data["rank"]
for key in data.keys():
print(key + " " + str(type(data[key].iloc[0])))
print(x.describe())
print(y.describe())
model = keras.Sequential(
[
layers.Dense(6, activation="relu"),
layers.Dense(10, activation="relu"),
layers.Dense(1),
]
)
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.25)
model.compile(
optimizer=keras.optimizers.RMSprop(),
loss=keras.losses.MeanSquaredError(),
metrics=[keras.metrics.MeanSquaredError()],
)
print(model.summary())
history = model.fit(
x_train, y_train, batch_size=64, epochs=2, validation_data=(x_test, y_test)
)
print(model.summary())
print(history.history)

# 
district <class 'str'>
district_o <class 'str'>
rank <class 'numpy.float64'>
set <class 'numpy.int64'>
area <class 'numpy.float64'>
perimeter <class 'numpy.float64'>
con_hull <class 'numpy.float64'>
reock <class 'numpy.float64'>
len_width <class 'numpy.float64'>
polsby_pop <class 'numpy.float64'>
geometry <class 'shapely.geometry.polygon.Polygon'>
               area     perimeter    con_hull       reock     len_width  polsby_pop
count  5.810000e+02  5.810000e+02  581.000000  581.000000  5.810000e+02  581.000000
mean   3.021031e+09  2.920331e+05    0.718060    0.381829  1.104009e+10    0.291362
std    1.553240e+10  8.141454e+05    0.136560    0.117522  1.157571e+11    0.140878
min    2.374961e+05  5.353782e+03    0.000028    0.000007  4.203100e+06    0.002712
25%    6.007942e+07  5.396089e+04    0.653352    0.304541  1.365093e+08    0.191961
50%    3.169981e+08  1.299022e+05    0.732183    0.385820  7.353186e+08    0.277472
75%    1.766022e+09  3.259159e+05    0.816594    0.467184  3.670517e+09    0.389346
max    3.316473e+11  1.478830e+07    0.989937    0.683143  2.563173e+12    0.768138
count    581.000000
mean      50.598461
std       28.581404
min        0.980392
25%       26.470588
50%       50.980392
75%       75.000000
max      100.000000
Name: rank, dtype: float64
Model: "sequential"
#

output:```┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓
┃ Layer (type) ┃ Output Shape ┃ Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩
│ dense (Dense) │ ? │ 0 (unbuilt) │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_1 (Dense) │ ? │ 0 (unbuilt) │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_2 (Dense) │ ? │ 0 (unbuilt) │
└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘
Total params: 0 (0.00 B)
Trainable params: 0 (0.00 B)
Non-trainable params: 0 (0.00 B)
None

# 
7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 13ms/step - loss: 659777345409777664000.0000 - mean_squared_error: 659777345409777664000.0000 - val_loss: 286636260213296988160.0000 - val_mean_squared_error: 286636260213296988160.0000
Epoch 2/2
7/7 ━━━━━━━━━━━━━━━━━━━━ 0s 3ms/step - loss: 1966408534886128287744.0000 - mean_squared_error: 1966408534886128287744.0000 - val_loss: 285890632999990460416.0000 - val_mean_squared_error: 285890632999990460416.0000
Model: "sequential"
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓
┃ Layer (type)                         ┃ Output Shape                ┃         Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩
│ dense (Dense)                        │ (None, 6)                   │              42 │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_1 (Dense)                      │ (None, 10)                  │              70 │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_2 (Dense)                      │ (None, 1)                   │              11 │
└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘
 Total params: 248 (996.00 B)
 Trainable params: 123 (492.00 B)
 Non-trainable params: 0 (0.00 B)
 Optimizer params: 125 (504.00 B)
None
{'loss': [8.465635066362714e+20, 8.465188928524627e+20], 'mean_squared_error': [8.465635066362714e+20, 8.465188928524627e+20], 'val_loss': [2.86636260213297e+20, 2.8589063299999046e+20], 'val_mean_squared_error': [2.86636260213297e+20, 2.8589063299999046e+20]}```
#

as you can see, all params are nan

#

idk why

#

there are no nans in the dataset

narrow bluffBOT
#

@supple laurel

Python help channel closed

This help channel has been closed and it's no longer possible to send messages here. If your question wasn't answered, feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.