Created a language model using LSTM Stack Dataset was from IBM for the training part
CREATION:-
- Hyperparameter description
- Tensorflow.Graph used
- Embedding vocabulary and converting to vectors
- Used 2 layers of LSTM stacked up together
- Output of 1st layer of LSTM would be one of the input of other layer
- Optimizer is to be used
- Class named PTBmodel is the most important part in model as everything is compiled over here
- At last the training and validation part
- The lesser the perplexeity, the more confident your model is 10)Perplexeity is randomness count so the above statement is valid.
DEPLOYMENT It is not done yet but one can deploy using several softwares line DL studio, etc THANK YOU Happy Coding!!!