-1

The problem is i am unable to find seq2seq in new tensorflow library

Here's the code-

decoderOutputs, states = tf.legacy_seq2seq.embedding_rnn_seq2seq( self.encoderInputs, # List<[batch=?, inputDim=1]>, list of size args.maxLength self.decoderInputs, # For training, we force the correct output (feed_previous=False) encoDecoCell, self.textData.getVocabularySize(), self.textData.getVocabularySize(), # Both encoder and decoder have the same number of class embedding_size=self.args.embeddingSize, # Dimension of each word output_projection=outputProjection.getWeights() if outputProjection else None, feed_previous=bool(self.args.test) # When we test (self.args.test), we use previous output as next input (feed_previous) )

    # training and reduce memory usage. Other solution, use sampling softmax

    # For testing only
    if self.args.test:
        if not outputProjection:
            self.outputs = decoderOutputs

        else:
            self.outputs = [outputProjection(output) for output in decoderOutputs]



    #For training only
    else:
        # Finally, we define the loss function
        self.lossFct = tf.compat.v1.legacy_seq2seq.sequence_loss(
            decoderOutputs,
            self.decoderTargets,
            self.decoderWeights,
            self.textData.getVocabularySize(),
            softmax_loss_function=sampledSoftmax if outputProjection else None  # If None, use default SoftMax
        )
        tf.summary.scalar('loss', self.lossFct)  # Keep track of the cost

        # Initialize the optimizerModel creation...
        opt = tf.train.AdamOptimizer(
            learning_rate=self.args.learningRate,
            beta1=0.9,
            beta2=0.999,
            epsilon=1e-08
        )
        self.optOp = opt.minimize(self.lossFct)

def step(self, batch):
    """ Forward/training step operation.
    Does not perform run on itself but just return the operators to do so. Those have then to be run
    Args:
        batch (Batch): Input data on testing mode, input and target on output mode
    Return:
        (ops), dict: A tuple of the (training, loss) operators or (outputs,) in testing mode with the associated feed dictionary
    """

    # Feed the dictionary
    feedDict = {}
    ops = None

    if not self.args.test:  # Training
        for i in range(self.args.maxLengthEnco):
            feedDict[self.encoderInputs[i]] = batch.encoderSeqs[i]
        for i in range(self.args.maxLengthDeco):
            feedDict[self.decoderInputs[i]] = batch.decoderSeqs[i]
            feedDict[self.decoderTargets[i]] = batch.targetSeqs[i]
            feedDict[self.decoderWeights[i]] = batch.weights[i]

        ops = (self.optOp, self.lossFct)
    else:  # Testing (batchSize == 1)
        for i in range(self.args.maxLengthEnco):
            feedDict[self.encoderInputs[i]] = batch.encoderSeqs[i]
        feedDict[self.decoderInputs[0]] = [self.textData.goToken]

        ops = (self.outputs,)

    # Return one pass operator
    return ops, feedDict

Problem showing to me-

module 'tensorflow' has no attribute 'legacy_seq2seq' File "C:\Users\krith\Downloads\Personality\Personality-Chatbot\chatbot\model.py", line 167, in buildNetwork decoderOutputs, states = tf.legacy_seq2seq.embedding_rnn_seq2seq( File "C:\Users\krith\Downloads\Personality\Personality-Chatbot\chatbot\model.py", line 105, in init self.buildNetwork() File "C:\Users\krith\Downloads\Personality\Personality-Chatbot\chatbot\chatbot.py", line 154, in main self.model = Model(self.args, self.textData) File "C:\Users\krith\Downloads\Personality\Personality-Chatbot\main.py", line 29, in chatbot.main()

Kritharth Singh
  • 1

1 Answers1

0

Seq2Seq (Sequence to sequence models) API can be find inside Tensorflow Addons. You need to run the code below to access Seq2Seq API and try running your code again:

!pip install tensorflow-addons
import tensorflow_addons as tfa
from tensorflow_addons import seq2seq

Please check this Seq2seq API link for more details in this and to access it's other functionalities.