Skip to content

By following the tutorial on TensorFlow , I generated a text using a character-based Recurrent Neural Networks. Dataset is based on Shakespeare's writing. Given a sequence of characters from the data, a LSTM-based model is trained to predict the next character in the sequence.

Notifications You must be signed in to change notification settings

smkatash/AI_projects

Repository files navigation

AI projects with TensorFlow

By following the tutorial on TensorFlow , I generated a text using a character-based Recurrent Neural Networks. Dataset is based on Shakespeare's writing. Given a sequence of characters from the data, a LSTM-based model is trained to predict the next character in the sequence. Longer sequences of text can be generated by calling the model repeatedly.

After applying temperature sampling with different temperature values, the following texts with conditional probability distributions over the vocabulary for the given temperature values were generated:

---------Temperature: 0.2---------
rs; a silly time
to make prescription for the fair strange:
the fortuly them the straight that the lands
and the will be our son the send thee head thee haste
that the day thee the will the was a good him thee death,
and the sting the commont a toomen and thee thee thee thee head thee will be some hands.

camillo:
the hand the day thee he
---------Temperature: 0.4---------
urselves: farewell.

earl of salisbury:
yet, their boldss of a mand best thee did come,
that thou hast thee sun wound thee hast head thee thee stays
the send thee what was a to the grace of thee
which shall pray thee with the king that have
the bear of this some son and thee here is thee art thee sund
you have stay thee say says the forth
---------Temperature: 0.6---------
.
montague:
and whosoe'er gainsays king, thee will subject.
the hand my father on thee than that dother,
or thou wilt them be some soul, every know,
him i be other tines of thing, thee set again;
with have stays of the will thee in thee.

camillo:
thou wast he not he do shall be in her consord,
and begent him that comes hoary of this lov
---------Temperature: 0.8---------
vereign liege was in my debt
upon remainsed or, but thee shall pointel!

york:
romeo's more, and son thee thou cousing not should be yet.

king henry vi:
being their leave to part you made thee,
and my wife the word manuth you think all he
now tull pity to contentor,
edward ton
and mantal have lea'e death lame a joys:
such breathing setit

About

By following the tutorial on TensorFlow , I generated a text using a character-based Recurrent Neural Networks. Dataset is based on Shakespeare's writing. Given a sequence of characters from the data, a LSTM-based model is trained to predict the next character in the sequence.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published