Styjun

How large would a mega structure have to be to host 1 billion people indefinitely?

What does it mean to "control target player"?

Why do textbooks often include the solutions to odd or even numbered problems but not both?

Trainee keeps missing deadlines for independent learning

Is "Busen" just the area between the breasts?

Why tighten down in a criss-cross pattern?

Why use cross notes in sheet music for hip hop tracks?

LWC - Best practice for routing?

Minimum distance between holes in inner tube

Is it illegal to withhold someone's passport and green card in California?

Parameterize chained calls to a utility program in Bash

Do I have to explain the mechanical superiority of the player-character within the fiction of the game?

How to avoid voltage drop when using full bridge rectifier as reverse polarity protection

What's currently blocking the construction of the wall between Mexico and the US?

Old sci-fi story: radiation mutated the animals, a boy loses a limb, but it's okay because "humans used to do great with only two arms"

How to make clear to people I don't want to answer their "Where are you from?" question?

What did River say when she woke from her proto-comatose state?

Why does Linux list NVMe drives as /dev/nvme0 instead of /dev/sda?

What was the Shuttle Carrier Aircraft escape tunnel?

What exactly is the 'online' in OLAP and OLTP?

Why did pressing the joystick button spit out keypresses?

How to remove this component from PCB

Why don't countries like Japan just print more money?

Is it damaging to turn off a small fridge for two days every week?

ValueError: Error when checking input: expected embedding_13_input to have 2 dimensions, but got array with shape (1, 1, 0)

Error when checking model input: expected lstm_1_input to have 3 dimensions, but got array with shape (339732, 29)How to use Keras LSTM with word embeddings to predict word id'sValueError: Error when checking input: expected input_1 to have shape (None, 1) but got array with shape (5, 54)KERAS: Get a SLICE of RNN timesteps with return_sequence = TrueValueError: Error when checking target: expected dense_3 to have 2 dimensions, but got array with shape (500, 10, 14)Keras LSTM from for loop, using functional API with custom number of layersValueError: Error when checking input: expected lstm_1_input to have shape (973, 215) but got array with shape (61, 215)ValueError: Error when checking target: expected dense_19 to have 3 dimensions, but got array with shape (5, 3)Keras/Tensorflow Input to RNN layersValueError: Error when checking input: expected lstm_1_input to have 3 dimensions, but got array with shape (6782, 36)

.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty height:90px;width:728px;box-sizing:border-box;

0

My code looks like this:

lr = 1e-3
window_length = 1
emb_size = 100
look_back = 10

expert_model = Sequential()
expert_model.add(Embedding(num_classes + 1, emb_size, input_length=look_back,mask_zero=True))
expert_model.add(LSTM(64, input_shape=(look_back,window_length)))
expert_model.add(Dense(num_classes, activation='softmax'))

All I want is to pass a list of classes of size 10 to an embedding layer and then to an LSTM one to predict the next class to come. Maybe the length of that list is not 10, so I put the mask_zero attribute to True and the vocabulary of the embedding layer with one extra value. Is this correct?

In addition, I'm not very sure what does the window_length means. Does it means the number of sequences to pass to the embedding? When I try to run this I get this error:

ValueError: Error when checking input: expected embedding_13_input to have 2 dimensions, but got array with shape (1, 1, 0)

To preprocess data I'm using a Processor object as this model is for a OpenAI environment called "RecoGym". The class is as follows:

class RecoProcessor(Processor):
 def process_observation(self, observation):
 if observation is None:
 return np.array([], dtype='float32')
 return np.array(observation, dtype='float32')

 def process_state_batch(self, batch):
 return np.array(batch).astype('float32')

 def process_reward(self, reward):
 return np.array(reward).astype('float32')

 def process_demo_data(self, demo_data):
 for step in demo_data:
 step[0] = self.process_observation(step[0])
 step[2] = self.process_reward(step[2])
 return demo_data

Please, I need some help. If you could only give me a tutorial on this I would be very grateful.

python keras lstm openai-gym

share|improve this question

edited Mar 25 at 9:56

Angelo

asked Mar 25 at 8:45

AngeloAngelo

16113

• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  
  Please also post the code you have used for preprocessing your data. There is a shape mismatch happening at preprocessing step.
  
  – user110327
  Mar 25 at 9:52
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @user110327 There you have. It might be a little bit strange, but OpenAI library requires that way. You can find the environment easily looking for RecoGym at Google if you need to.
  
  – Angelo
  Mar 25 at 9:58
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  Window size is the length of the sequence. For example, if your sequence is "1,2,3.4" then a window size of 2 gives "1,2", "2,3", "3,4". I am sorry I am unable to help you with the actual issue because the input shape is still not clear to me.
  
  – user110327
  Mar 25 at 10:04
  

  
  
  
  

add a comment | 

0

My code looks like this:

lr = 1e-3
window_length = 1
emb_size = 100
look_back = 10

expert_model = Sequential()
expert_model.add(Embedding(num_classes + 1, emb_size, input_length=look_back,mask_zero=True))
expert_model.add(LSTM(64, input_shape=(look_back,window_length)))
expert_model.add(Dense(num_classes, activation='softmax'))

All I want is to pass a list of classes of size 10 to an embedding layer and then to an LSTM one to predict the next class to come. Maybe the length of that list is not 10, so I put the mask_zero attribute to True and the vocabulary of the embedding layer with one extra value. Is this correct?

In addition, I'm not very sure what does the window_length means. Does it means the number of sequences to pass to the embedding? When I try to run this I get this error:

ValueError: Error when checking input: expected embedding_13_input to have 2 dimensions, but got array with shape (1, 1, 0)

To preprocess data I'm using a Processor object as this model is for a OpenAI environment called "RecoGym". The class is as follows:

class RecoProcessor(Processor):
 def process_observation(self, observation):
 if observation is None:
 return np.array([], dtype='float32')
 return np.array(observation, dtype='float32')

 def process_state_batch(self, batch):
 return np.array(batch).astype('float32')

 def process_reward(self, reward):
 return np.array(reward).astype('float32')

 def process_demo_data(self, demo_data):
 for step in demo_data:
 step[0] = self.process_observation(step[0])
 step[2] = self.process_reward(step[2])
 return demo_data

Please, I need some help. If you could only give me a tutorial on this I would be very grateful.

python keras lstm openai-gym

share|improve this question

edited Mar 25 at 9:56

Angelo

asked Mar 25 at 8:45

AngeloAngelo

16113

• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  
  Please also post the code you have used for preprocessing your data. There is a shape mismatch happening at preprocessing step.
  
  – user110327
  Mar 25 at 9:52
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @user110327 There you have. It might be a little bit strange, but OpenAI library requires that way. You can find the environment easily looking for RecoGym at Google if you need to.
  
  – Angelo
  Mar 25 at 9:58
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  Window size is the length of the sequence. For example, if your sequence is "1,2,3.4" then a window size of 2 gives "1,2", "2,3", "3,4". I am sorry I am unable to help you with the actual issue because the input shape is still not clear to me.
  
  – user110327
  Mar 25 at 10:04
  

  
  
  
  

add a comment | 

My code looks like this:

lr = 1e-3
window_length = 1
emb_size = 100
look_back = 10

expert_model = Sequential()
expert_model.add(Embedding(num_classes + 1, emb_size, input_length=look_back,mask_zero=True))
expert_model.add(LSTM(64, input_shape=(look_back,window_length)))
expert_model.add(Dense(num_classes, activation='softmax'))

All I want is to pass a list of classes of size 10 to an embedding layer and then to an LSTM one to predict the next class to come. Maybe the length of that list is not 10, so I put the mask_zero attribute to True and the vocabulary of the embedding layer with one extra value. Is this correct?

In addition, I'm not very sure what does the window_length means. Does it means the number of sequences to pass to the embedding? When I try to run this I get this error:

ValueError: Error when checking input: expected embedding_13_input to have 2 dimensions, but got array with shape (1, 1, 0)

To preprocess data I'm using a Processor object as this model is for a OpenAI environment called "RecoGym". The class is as follows:

class RecoProcessor(Processor):
 def process_observation(self, observation):
 if observation is None:
 return np.array([], dtype='float32')
 return np.array(observation, dtype='float32')

 def process_state_batch(self, batch):
 return np.array(batch).astype('float32')

 def process_reward(self, reward):
 return np.array(reward).astype('float32')

 def process_demo_data(self, demo_data):
 for step in demo_data:
 step[0] = self.process_observation(step[0])
 step[2] = self.process_reward(step[2])
 return demo_data

Please, I need some help. If you could only give me a tutorial on this I would be very grateful.

python keras lstm openai-gym

share|improve this question

edited Mar 25 at 9:56

Angelo

asked Mar 25 at 8:45

AngeloAngelo

16113

lr = 1e-3
window_length = 1
emb_size = 100
look_back = 10

expert_model = Sequential()
expert_model.add(Embedding(num_classes + 1, emb_size, input_length=look_back,mask_zero=True))
expert_model.add(LSTM(64, input_shape=(look_back,window_length)))
expert_model.add(Dense(num_classes, activation='softmax'))

ValueError: Error when checking input: expected embedding_13_input to have 2 dimensions, but got array with shape (1, 1, 0)

class RecoProcessor(Processor):
 def process_observation(self, observation):
 if observation is None:
 return np.array([], dtype='float32')
 return np.array(observation, dtype='float32')

 def process_state_batch(self, batch):
 return np.array(batch).astype('float32')

 def process_reward(self, reward):
 return np.array(reward).astype('float32')

 def process_demo_data(self, demo_data):
 for step in demo_data:
 step[0] = self.process_observation(step[0])
 step[2] = self.process_reward(step[2])
 return demo_data

share|improve this question

edited Mar 25 at 9:56

Angelo

asked Mar 25 at 8:45

AngeloAngelo

16113

share|improve this question

edited Mar 25 at 9:56

Angelo

asked Mar 25 at 8:45

AngeloAngelo

16113

asked Mar 25 at 8:45

AngeloAngelo

16113

asked Mar 25 at 8:45

AngeloAngelo

16113