How to input different sized images into transfer learning network The Next CEO of Stack Overflow2019 Community Moderator ElectionTensorflow oscillating Test and Train Accuracy?Accuracy drops if more layers trainable - weirdFine tuning accuracy lower than Raw Transfer Learning AccuracyInterpreting confusion matrix and validation results in convolutional networksHow to improve loss and avoid overfittingDifficulty in choosing Hyperparameters for my CNNHow to set input for proper fit with lstm?Multi-label classification, recall and precision increase but accuracy decrease, why?Using deep learning to classify similar imagesHow to properly resize input images for transfer learning
What did we know about the Kessel run before the prologues?
What happened in Rome, when the western empire "fell"?
Help understanding this unsettling image of Titan, Epimetheus, and Saturn's rings?
Make solar eclipses exceedingly rare, but still have new moons
Newlines in BSD sed vs gsed
Measuring resistivity of dielectric liquid
A Man With a Stainless Steel Endoskeleton (like The Terminator) Fighting Cloaked Aliens Only He Can See
Domestic-to-international connection at Orlando (MCO)
Do I need to write [sic] when a number is less than 10 but isn't written out?
Is this "being" usage is essential?
Can MTA send mail via a relay without being told so?
Math-accent symbol over parentheses enclosing accented symbol (amsmath)
Minecraft Executing if more than 500 entities
If the updated MCAS software needs two AOA sensors, doesn't that introduce a new single point of failure?
Why isn't the Mueller report being released completely and unredacted?
If Nick Fury and Coulson already knew about aliens (Kree and Skrull) why did they wait until Thor's appearance to start making weapons?
Running a General Election and the European Elections together
Received an invoice from my ex-employer billing me for training; how to handle?
How do I align (1) and (2)?
Proper way to express "He disappeared them"
Why don't programming languages automatically manage the synchronous/asynchronous problem?
If a black hole is created from light, can this black hole then move at the speed of light?
Why, when going from special to general relativity, do we just replace partial derivatives with covariant derivatives?
Why does the flight controls check come before arming the autobrake on the A320?
How to input different sized images into transfer learning network
The Next CEO of Stack Overflow2019 Community Moderator ElectionTensorflow oscillating Test and Train Accuracy?Accuracy drops if more layers trainable - weirdFine tuning accuracy lower than Raw Transfer Learning AccuracyInterpreting confusion matrix and validation results in convolutional networksHow to improve loss and avoid overfittingDifficulty in choosing Hyperparameters for my CNNHow to set input for proper fit with lstm?Multi-label classification, recall and precision increase but accuracy decrease, why?Using deep learning to classify similar imagesHow to properly resize input images for transfer learning
$begingroup$
I have been looking online for a solution but have a difficult time finding a clear enough solution. I want to know how to use transfer learning (VGG16 for example) on images that have different sizes than the images the network originally trained on (so instead of inputting images of size (224,224,3) I want to input images of size (32,32,3)).
I initially thought about just padding those images but the network may look into the black pixels and think that they mean something, and I realize that might hard the accuracy and also when I tried to do that my colab notebook collapsed.
This is my VGG-16 code:
def vgg16_model(img_rows, img_cols, channel=1, num_classes=None):
model = VGG16(weights='imagenet', include_top=True)
model.layers.pop()
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes = []
x=Dense(num_classes, activation='relu')(model.output)
model=Model(model.input,x)
#To set the first 8 layers to non-trainable (weights will not be updated)
for layer in model.layers[:15]:
layer.trainable = False
for layer in model.layers[16:]:
layer.trainable=True
model_new = Sequential()
for layer in model.layers[:-1]: # just exclude last layer from copying
model_new.add(layer)
model=model_new
model.add(Dense(256,activation='relu',input_shape=(1000,)))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(10,activation='softmax'))
#model.add(Dense(1,activation='softmax'))
# Learning rate is changed to 0.001
#sgd = SGD(lr=1e-2, decay=1e-6, momentum=0.9, nesterov=True)
sgd = SGD(lr=lr,decay=decay,momentum=0.95, nesterov=True)
adam=Adam(lr=1e-5, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0001, amsgrad=True)
#model.compile(optimizer=adam, loss='binary_crossentropy',metrics=['accuracy'])
model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])
# checkpoint
filepath="weights-improvement-epoch:02d-val_acc:.2f.hdf5"
checkpoint = ModelCheckpoint(filepath, monitor='val_acc', verbose=1, save_best_only=True, mode='max')
callbacks_list = [checkpoint]
return model
I would greatly appreciate if someone could write the corrected version of this code in order to enable me to insert pictures of size (32,32,3).
Thanks a lot in advance!!
deep-learning transfer-learning
asked Mar 23 at 16:16
KerenKeren
262
$endgroup$
add a comment |
$begingroup$
I have been looking online for a solution but have a difficult time finding a clear enough solution. I want to know how to use transfer learning (VGG16 for example) on images that have different sizes than the images the network originally trained on (so instead of inputting images of size (224,224,3) I want to input images of size (32,32,3)).
I initially thought about just padding those images but the network may look into the black pixels and think that they mean something, and I realize that might hard the accuracy and also when I tried to do that my colab notebook collapsed.
This is my VGG-16 code:
def vgg16_model(img_rows, img_cols, channel=1, num_classes=None):
model = VGG16(weights='imagenet', include_top=True)
model.layers.pop()
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes = []
x=Dense(num_classes, activation='relu')(model.output)
model=Model(model.input,x)
#To set the first 8 layers to non-trainable (weights will not be updated)
for layer in model.layers[:15]:
layer.trainable = False
for layer in model.layers[16:]:
layer.trainable=True
model_new = Sequential()
for layer in model.layers[:-1]: # just exclude last layer from copying
model_new.add(layer)
model=model_new
model.add(Dense(256,activation='relu',input_shape=(1000,)))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(10,activation='softmax'))
#model.add(Dense(1,activation='softmax'))
# Learning rate is changed to 0.001
#sgd = SGD(lr=1e-2, decay=1e-6, momentum=0.9, nesterov=True)
sgd = SGD(lr=lr,decay=decay,momentum=0.95, nesterov=True)
adam=Adam(lr=1e-5, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0001, amsgrad=True)
#model.compile(optimizer=adam, loss='binary_crossentropy',metrics=['accuracy'])
model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])
# checkpoint
filepath="weights-improvement-epoch:02d-val_acc:.2f.hdf5"
checkpoint = ModelCheckpoint(filepath, monitor='val_acc', verbose=1, save_best_only=True, mode='max')
callbacks_list = [checkpoint]
return model
I would greatly appreciate if someone could write the corrected version of this code in order to enable me to insert pictures of size (32,32,3).
Thanks a lot in advance!!
deep-learning transfer-learning
asked Mar 23 at 16:16
KerenKeren
262
$endgroup$
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59
add a comment |
$begingroup$
I have been looking online for a solution but have a difficult time finding a clear enough solution. I want to know how to use transfer learning (VGG16 for example) on images that have different sizes than the images the network originally trained on (so instead of inputting images of size (224,224,3) I want to input images of size (32,32,3)).
I initially thought about just padding those images but the network may look into the black pixels and think that they mean something, and I realize that might hard the accuracy and also when I tried to do that my colab notebook collapsed.
This is my VGG-16 code:
def vgg16_model(img_rows, img_cols, channel=1, num_classes=None):
model = VGG16(weights='imagenet', include_top=True)
model.layers.pop()
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes = []
x=Dense(num_classes, activation='relu')(model.output)
model=Model(model.input,x)
#To set the first 8 layers to non-trainable (weights will not be updated)
for layer in model.layers[:15]:
layer.trainable = False
for layer in model.layers[16:]:
layer.trainable=True
model_new = Sequential()
for layer in model.layers[:-1]: # just exclude last layer from copying
model_new.add(layer)
model=model_new
model.add(Dense(256,activation='relu',input_shape=(1000,)))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(10,activation='softmax'))
#model.add(Dense(1,activation='softmax'))
# Learning rate is changed to 0.001
#sgd = SGD(lr=1e-2, decay=1e-6, momentum=0.9, nesterov=True)
sgd = SGD(lr=lr,decay=decay,momentum=0.95, nesterov=True)
adam=Adam(lr=1e-5, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0001, amsgrad=True)
#model.compile(optimizer=adam, loss='binary_crossentropy',metrics=['accuracy'])
model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])
# checkpoint
filepath="weights-improvement-epoch:02d-val_acc:.2f.hdf5"
checkpoint = ModelCheckpoint(filepath, monitor='val_acc', verbose=1, save_best_only=True, mode='max')
callbacks_list = [checkpoint]
return model
I would greatly appreciate if someone could write the corrected version of this code in order to enable me to insert pictures of size (32,32,3).
Thanks a lot in advance!!
deep-learning transfer-learning
asked Mar 23 at 16:16
KerenKeren
262
$endgroup$
I have been looking online for a solution but have a difficult time finding a clear enough solution. I want to know how to use transfer learning (VGG16 for example) on images that have different sizes than the images the network originally trained on (so instead of inputting images of size (224,224,3) I want to input images of size (32,32,3)).
I initially thought about just padding those images but the network may look into the black pixels and think that they mean something, and I realize that might hard the accuracy and also when I tried to do that my colab notebook collapsed.
This is my VGG-16 code:
def vgg16_model(img_rows, img_cols, channel=1, num_classes=None):
model = VGG16(weights='imagenet', include_top=True)
model.layers.pop()
model.outputs = [model.layers[-1].output]
model.layers[-1].outbound_nodes = []
x=Dense(num_classes, activation='relu')(model.output)
model=Model(model.input,x)
#To set the first 8 layers to non-trainable (weights will not be updated)
for layer in model.layers[:15]:
layer.trainable = False
for layer in model.layers[16:]:
layer.trainable=True
model_new = Sequential()
for layer in model.layers[:-1]: # just exclude last layer from copying
model_new.add(layer)
model=model_new
model.add(Dense(256,activation='relu',input_shape=(1000,)))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(256,activation='relu'))
model.add(Dropout(0.3))
model.add(Dense(10,activation='softmax'))
#model.add(Dense(1,activation='softmax'))
# Learning rate is changed to 0.001
#sgd = SGD(lr=1e-2, decay=1e-6, momentum=0.9, nesterov=True)
sgd = SGD(lr=lr,decay=decay,momentum=0.95, nesterov=True)
adam=Adam(lr=1e-5, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0001, amsgrad=True)
#model.compile(optimizer=adam, loss='binary_crossentropy',metrics=['accuracy'])
model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])
# checkpoint
filepath="weights-improvement-epoch:02d-val_acc:.2f.hdf5"
checkpoint = ModelCheckpoint(filepath, monitor='val_acc', verbose=1, save_best_only=True, mode='max')
callbacks_list = [checkpoint]
return model
I would greatly appreciate if someone could write the corrected version of this code in order to enable me to insert pictures of size (32,32,3).
Thanks a lot in advance!!
deep-learning transfer-learning
deep-learning transfer-learning
asked Mar 23 at 16:16
KerenKeren
262
asked Mar 23 at 16:16
KerenKeren
262
asked Mar 23 at 16:16
KerenKeren
262
asked Mar 23 at 16:16
KerenKeren
262
asked Mar 23 at 16:16
KerenKeren
262
262
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59
add a comment |
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59
add a comment |
1 Answer
1
active
oldest
votes
$begingroup$
Resizing is the best option, if they are bigger downscale them, else upscale them.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
add a comment |
Your Answer
StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "557"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f47851%2fhow-to-input-different-sized-images-into-transfer-learning-network%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
Resizing is the best option, if they are bigger downscale them, else upscale them.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
add a comment |
$begingroup$
Resizing is the best option, if they are bigger downscale them, else upscale them.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
add a comment |
$begingroup$
Resizing is the best option, if they are bigger downscale them, else upscale them.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
Resizing is the best option, if they are bigger downscale them, else upscale them.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
answered Mar 24 at 15:29
Amita KapoorAmita Kapoor
111
111
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
New contributor
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
Amita Kapoor is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
add a comment |
add a comment |
Thanks for contributing an answer to Data Science Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f47851%2fhow-to-input-different-sized-images-into-transfer-learning-network%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
$begingroup$
Have you tried upscaling images to 224*224 ?
$endgroup$
– Shamit Verma
Mar 23 at 16:39
$begingroup$
Yeah, I had a hard time finding an elegant, simple way of doing that but I ended up figuring it out, thanks!
$endgroup$
– Keren
Mar 24 at 19:59