'NoneType' object has no attribute '_inbound_nodes' ? Keras seq2seq classification
I have a problem with creating a Keras model. I found a simple encoder decoder and try to fix it like below:
# some encoder code ... .... above is not shown here, where it is too obvious
encoder_outputs, state_h, state_c = encoder(encoder_inputs)
encoder_states = [state_h, state_c]
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, _, _ = decoder_lstm(encoder_outputs[-1:], initial_state=encoder_states)
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
which will use the last encoder output as decoder input and only goes for a single output.
I wonder why it creates such a problem at:
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
and creating message:
'NoneType' object has no attribute '_inbound_nodes'
How do I solve it? I tried to find similar questions' answers but I didn't get a good one to fix mine.
python machine-learning keras classification lstm
edited Nov 23 '18 at 12:11
today
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
add a comment |
I have a problem with creating a Keras model. I found a simple encoder decoder and try to fix it like below:
# some encoder code ... .... above is not shown here, where it is too obvious
encoder_outputs, state_h, state_c = encoder(encoder_inputs)
encoder_states = [state_h, state_c]
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, _, _ = decoder_lstm(encoder_outputs[-1:], initial_state=encoder_states)
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
which will use the last encoder output as decoder input and only goes for a single output.
I wonder why it creates such a problem at:
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
and creating message:
'NoneType' object has no attribute '_inbound_nodes'
How do I solve it? I tried to find similar questions' answers but I didn't get a good one to fix mine.
python machine-learning keras classification lstm
edited Nov 23 '18 at 12:11
today
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
add a comment |
I have a problem with creating a Keras model. I found a simple encoder decoder and try to fix it like below:
# some encoder code ... .... above is not shown here, where it is too obvious
encoder_outputs, state_h, state_c = encoder(encoder_inputs)
encoder_states = [state_h, state_c]
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, _, _ = decoder_lstm(encoder_outputs[-1:], initial_state=encoder_states)
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
which will use the last encoder output as decoder input and only goes for a single output.
I wonder why it creates such a problem at:
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
and creating message:
'NoneType' object has no attribute '_inbound_nodes'
How do I solve it? I tried to find similar questions' answers but I didn't get a good one to fix mine.
python machine-learning keras classification lstm
edited Nov 23 '18 at 12:11
today
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
I have a problem with creating a Keras model. I found a simple encoder decoder and try to fix it like below:
# some encoder code ... .... above is not shown here, where it is too obvious
encoder_outputs, state_h, state_c = encoder(encoder_inputs)
encoder_states = [state_h, state_c]
decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)
decoder_outputs, _, _ = decoder_lstm(encoder_outputs[-1:], initial_state=encoder_states)
decoder_dense = Dense(num_decoder_tokens, activation='softmax')
decoder_outputs = decoder_dense(decoder_outputs)
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
which will use the last encoder output as decoder input and only goes for a single output.
I wonder why it creates such a problem at:
model = Model(inputs=[encoder_inputs], outputs=decoder_outputs)
and creating message:
'NoneType' object has no attribute '_inbound_nodes'
How do I solve it? I tried to find similar questions' answers but I didn't get a good one to fix mine.
python machine-learning keras classification lstm
python machine-learning keras classification lstm
edited Nov 23 '18 at 12:11
today
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
edited Nov 23 '18 at 12:11
today
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
edited Nov 23 '18 at 12:11
today
11.5k22239
edited Nov 23 '18 at 12:11
today
11.5k22239
edited Nov 23 '18 at 12:11
today
11.5k22239
11.5k22239
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
asked Nov 23 '18 at 9:35
Isaac SimIsaac Sim
10511
10511
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
First of all, encoder_outputs[-1:] would give you the last batch, not the last output of each batch which is encoder_outputs[:,-1:].
Second, since you need to pass Keras Tensors to layers in Keras, you need to use a Lambda layer to do the slicing:
last_input = Lambda(lambda x: x[:,-1:])(encoder_inputs)
decoder_outputs, _, _ = decoder_lstm(last_input,
initial_state=encoder_states)
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size usingbatch_input_shapeor it is implicitly added by Keras itself.
– today
Nov 26 '18 at 9:13
add a comment |
StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
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: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
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%2fstackoverflow.com%2fquestions%2f53443996%2fnonetype-object-has-no-attribute-inbound-nodes-keras-seq2seq-classificati%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
First of all, encoder_outputs[-1:] would give you the last batch, not the last output of each batch which is encoder_outputs[:,-1:].
Second, since you need to pass Keras Tensors to layers in Keras, you need to use a Lambda layer to do the slicing:
last_input = Lambda(lambda x: x[:,-1:])(encoder_inputs)
decoder_outputs, _, _ = decoder_lstm(last_input,
initial_state=encoder_states)
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size usingbatch_input_shapeor it is implicitly added by Keras itself.
– today
Nov 26 '18 at 9:13
add a comment |
First of all, encoder_outputs[-1:] would give you the last batch, not the last output of each batch which is encoder_outputs[:,-1:].
Second, since you need to pass Keras Tensors to layers in Keras, you need to use a Lambda layer to do the slicing:
last_input = Lambda(lambda x: x[:,-1:])(encoder_inputs)
decoder_outputs, _, _ = decoder_lstm(last_input,
initial_state=encoder_states)
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size usingbatch_input_shapeor it is implicitly added by Keras itself.
– today
Nov 26 '18 at 9:13
add a comment |
First of all, encoder_outputs[-1:] would give you the last batch, not the last output of each batch which is encoder_outputs[:,-1:].
Second, since you need to pass Keras Tensors to layers in Keras, you need to use a Lambda layer to do the slicing:
last_input = Lambda(lambda x: x[:,-1:])(encoder_inputs)
decoder_outputs, _, _ = decoder_lstm(last_input,
initial_state=encoder_states)
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
First of all, encoder_outputs[-1:] would give you the last batch, not the last output of each batch which is encoder_outputs[:,-1:].
Second, since you need to pass Keras Tensors to layers in Keras, you need to use a Lambda layer to do the slicing:
last_input = Lambda(lambda x: x[:,-1:])(encoder_inputs)
decoder_outputs, _, _ = decoder_lstm(last_input,
initial_state=encoder_states)
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
answered Nov 23 '18 at 12:09
todaytoday
11.5k22239
11.5k22239
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size usingbatch_input_shapeor it is implicitly added by Keras itself.
– today
Nov 26 '18 at 9:13
add a comment |
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size usingbatch_input_shapeor it is implicitly added by Keras itself.
– today
Nov 26 '18 at 9:13
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
well, sorry to not inform you that the input type of decoder [does not include] batch size. Input shape is [sequence length x feature length]. So your guess that encoder_outputs[-1:] is the last batch is wrong. But anyways, Lambda seems working although I have to test a few things more. Once I am confident with it, I will come back and accept your answer. Thanks
– Isaac Sim
Nov 26 '18 at 0:45
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size using
batch_input_shape or it is implicitly added by Keras itself.– today
Nov 26 '18 at 9:13
@IsaacSim Oh, my dear! It is impossible that the batch axis is not present in the tensors. Either you have explicitly set the batch size using
batch_input_shape or it is implicitly added by Keras itself.– today
Nov 26 '18 at 9:13
add a comment |
Thanks for contributing an answer to Stack Overflow!
- 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.
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%2fstackoverflow.com%2fquestions%2f53443996%2fnonetype-object-has-no-attribute-inbound-nodes-keras-seq2seq-classificati%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
