Generating Poetic Texts with Recurrent Neural Networks in Python

In this video we are creating our first AI project. It is a recurrent neural network that generates poetic texts, similar to those of Shakespeare!

Subscribe and Like for more free content!

[ananthramvijayaraj]

Bro, I'm surprised that such nice videos have very little views. Really love the work and just want to tell you to keep up the grind!!

[ahmetyusuf]

For all the newcommers in AI, quick tip: "lr" from the line “model.compile(loss=’categorical_crossentropy’, optimizer=RMSprop(lr=0.01))” has been chnaged to learning_rate in the newer versions. Enjoy!

[jackroul]

More than the coding itself I liked the way you explains the concept. Hope to see more😊

[RehaERMETin]

Just so people know in more recent versions of TensorFlow/Keras, the lr argument has been replaced with learning_rate

[matthewweitzner]

Great tutorial! The website did not connect properly for me, but I then switched it out with some of my own writing that I converted to a .txt file. Wasn't as polished, but then again I'm not as prolific a writer as Shakespeare, so it had less training data. If you're considering starting the video, I would highly recommend. Also, afterwards you can edit the function so that you can input your own starting text if you want.

[lynx-meew]

def sample(preds, temperature=1.0):
preds =
preds = np.log(preds) / temperature
exp_preds = np.exp(preds)
preds = exp_preds / np.sum(exp_preds)
probas = np.random.multinomial(1, preds, 1)
return np.argmax(probas)

[SoulFrmTitanic]

i saw you for the first time 8 years ago
Rn, I'm 20...

[mohamedaminelahlal]

Can you upload more python for networking and "ethical hacking"? Continue your videos are awesome

[mohorbanerjee]

Thanks for the amazing tutorial! But when I tried out with the exact same code, even for moderate temperatures like 0.6 and 0.8 I am getting very haphazardly generated results most of which doesn't really make any sense....may I know how do I obtain the refined results as were shown in the video for 0.6 and 0.8?

[mishahappy]

Really cool video. you deserve a lot of views

[sakari.niittymaa]

This is so good tutorial and excellent teacher! Thank you for your videos. +1 sub 👍

[Memories-Lens]

is there any other specific thing do we have to learn other than python for these projects?

[raphaelmgonja]

i countered this problem from you code on chatbot
ValueError Traceback (most recent call last)
Cell In[168], line 21
19 # Shuffle and convert to a NumPy array after the loop
20 random.shuffle(training)
---> 21 training = np.array(training)
23 train_x = list(training[:, 0])
24 train_y = list(training[:, 1])

ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 2 dimensions. The detected shape was (12, 2) + inhomogeneous part.

[techandtalks]

bro this line x[i, t, char_to_index[character]] = 1 is producing blunt, \nand rice ap thomas with a valian' error

[eldranoson]

How would you be able to make it create more senseful text?
Putting in all of the shakespear text instead of just a part of it?

[aneeskhann]

can someone name me the tool or framework used in this video to build the project

[arvis_]

Hey, nice tutorial, i do get ValueError: Input 0 of layer sequential not compatible with the layer on the predictions = model.predict

[jory_a]

Thank u ❤ i do my project in univerfity from your code

[sargamyadav]

Nice tutorial 😀. Could you please cover text generation for dialogue systems if possible?

[RobertLogic]

I just did this, I had to swap some things around to make it work though.
How hard would it be to query this model like a bot?
Is there something more specific to this that I can look up?