Train on full dataset about nolearn HOT 5 CLOSED

Thing is - when do you stop ? when you train on full data set.

from nolearn.

Also I am getting an error when I train on full data set.
/usr/lib/python2.7/dist-packages/numpy/core/_methods.py:55: RuntimeWarning: Mean of empty slice.
I only modified the train_test_split method in lasagne.py.

from nolearn.

I have tried this out and after it runs I get the error:

/usr/lib64/python2.7/site-packages/numpy/core/methods.py:55: RuntimeWarning: Mean of empty slice.
warnings.warn("Mean of empty slice.", RuntimeWarning)
1 | 3.379510 | nan | nan | nan% | 267.6s
2 | 2.429299 | nan | nan | nan% | 267.7s
3 | 2.071116 | nan | nan | nan% | 267.4s
4 | 1.854738 | nan | nan | nan% | 267.3s
5 | 1.683506 | nan | nan | nan% | 267.1s
6 | 1.568097 | nan | nan | nan% | 267.1s
7 | 1.478959 | nan | nan | nan% | 267.1s
8 | 1.407118 | nan | nan | nan% | 267.0s
.......
Early stopping.
Best valid loss was inf at epoch 0.
Traceback (most recent call last):
File "run_analysis.py", line 121, in
func(*sys.argv[2:])
File "run_analysis.py", line 109, in fit
model.fit(X, y)
File "/home/segalam/TEST/src/nolearn/nolearn/lasagne.py", line 244, in fit
self.train_loop(X, y)
File "/home/segalam/TEST/src/nolearn/nolearn/lasagne.py", line 326, in train_loop
func(self, self.train_history)
File "/home/segalam/TEST/src/lasagne/examples/helpers.py", line 277, in call
nn.load_weights_from(self.best_weights)
File "/home/segalam/TEST/src/nolearn/nolearn/lasagne.py", line 384, in load_weights_from
w.get_value() if hasattr(w, 'get_value') else w for w in source]
TypeError: 'NoneType' object is not iterable

from nolearn.

You should not use Early stopping if you have no evaluation set, since early stopping compares the costs on the evaluation set.

from nolearn.

indeed, that fixes the error. Thanks!

from nolearn.