lunny/keras
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add simpler verbose mode to Sequential model

Browse Source
This commit is contained in:
fchollet 2015-04-16 15:05:59 -07:00
parent 5f25e2205a
commit 731f0ab42b
1 changed files with 29 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
keras/models.py
View File

@ -121,7 +121,8 @@ class Sequential(object):
np.random.shuffle(index_array)
batches = make_batches(len(X), batch_size)
progbar = Progbar(target=len(X))
if verbose==1:
progbar = Progbar(target=len(X))
for batch_index, (batch_start, batch_end) in enumerate(batches):
if shuffle:
batch_ids = index_array[batch_start:batch_end]
@ -138,23 +139,30 @@ class Sequential(object):
# logging
if verbose:
is_last_batch = (batch_index == len(batches) - 1)
if not is_last_batch or not do_validation:
if show_accuracy:
progbar.update(batch_end, [('loss', loss), ('acc.', acc)])
else:
progbar.update(batch_end, [('loss', loss)])
if (not is_last_batch or not do_validation):
if verbose==1:
if show_accuracy:
progbar.update(batch_end, [('loss', loss), ('acc.', acc)])
else:
progbar.update(batch_end, [('loss', loss)])
else:
if show_accuracy:
val_loss, val_acc = self.test(X_val, y_val, accuracy=True)
progbar.update(batch_end, [('loss', loss), ('acc.', acc), ('val. loss', val_loss), ('val. acc.', val_acc)])
if verbose==1:
progbar.update(batch_end, [('loss', loss), ('acc.', acc), ('val. loss', val_loss), ('val. acc.', val_acc)])
if verbose==2:
print("loss: %.4f - acc.: %.4f - val. loss: %.4f - val. acc.: %.4f" % (loss, acc, val_loss, val_acc))
else:
val_loss = self.test(X_val, y_val, accuracy=False)
progbar.update(batch_end, [('loss', loss), ('val. loss', val_loss)])
if verbose==1:
progbar.update(batch_end, [('loss', loss), ('val. loss', val_loss)])
if verbose==2:
print("loss: %.4f - acc.: %.4f" % (loss, acc))
def predict_proba(self, X, batch_size=128, verbose=1):
batches = make_batches(len(X), batch_size)
if verbose:
if verbose==1:
progbar = Progbar(target=len(X))
for batch_index, (batch_start, batch_end) in enumerate(batches):
X_batch = X[batch_start:batch_end]
@ -165,8 +173,9 @@ class Sequential(object):
preds = np.zeros(shape)
preds[batch_start:batch_end] = batch_preds
if verbose:
if verbose==1:
progbar.update(batch_end)
return preds
@ -199,10 +208,16 @@ class Sequential(object):
tot_score += loss
if verbose:
if show_accuracy:
progbar.update(batch_end, [('loss', loss), ('acc.', acc)])
else:
progbar.update(batch_end, [('loss', loss)])
if verbose==1:
if show_accuracy:
progbar.update(batch_end, [('loss', loss), ('acc.', acc)])
else:
progbar.update(batch_end, [('loss', loss)])
if batch_index == len(batches) and verbose==2:
if show_accuracy:
print("loss: %.4f - acc.: %.4f" % (loss, acc))
else:
print("loss: %.4f")
if show_accuracy:
return tot_score/len(batches), tot_acc/len(batches)