Neural network tuning in Python|Hyper parameter tuning of neural network in python

Neural network tuning in Python|Hyper parameter tuning of neural network in python
#NeuralNetworkTuningPython #NeuralNetworkPython

Welcome! I'm Aman, a Data Scientist & AI Mentor.

🚀 **Level Up Your Skills:**

* **Udemy Courses:** 🔥 **Start Learning Now!** 🔥
* X (Twitter): @unfolds
* Instagram: @unfold_data_science

🎬 **Featured Playlists:**

🎥 **My Studio Gear:**

#DataScience #AI #MachineLearning

About this video:
In this video, I talk about tuning a neural network in Python.
I explain, how different model parameters can be tuned in python using gridsearchcv.
Below questions are answered in this video
1. How to tune neural network in python
2. Neural network hyper parameter tuning in python
3. How to tune weight in neural network?
4. How to tune epoch and batch size in neural network?
5. How to tune optimizer in neural network?
6. Python Neural network hyper parameter tuning

[srijitaacharyya]

Thank you so very much sir. I have always found understanding Neural Network a difficult job. For months I have searched different articles and courses but failed. Your videos explained the theoretical as well as the application parts brilliantly. I am now confident to practice on my own. 😊😊

[praveenkuthuru]

Excellent work Aman....keep doing such videos, impressive...

[AkaExcel]

Thank You, would be great if you can make full video with all of the above hyperparams! Best wishes!

[الدينوالحياة-خي]

Thanks, I need to understand keras tuner optimization for datasets csv

[jeronsout]

Sir, Maybe You have already explained but I'm looking for same tutorial for continues (target variable) output not binary target variable(like this tutorial). specifically for example predict the house price based on room numbers, address, and so on. I'm interested to do it by RNN and LSTM. I'll be happy and appreciate if you explain or refer me for this matter since I've found your explanations so clear.

[sandipansarkar]

great explanation. Need to practice in colab

[DS_AIML]

I verified my Neural Network model parameters and this is what best i could get for below parameters
'batch_size': 10, 'epochs': 50, activation': 'tanh', 'neurons': 15, 'optimizer': 'Adagrad'

[liweiwang]

Hi sir, when running part 1, each time I got a different bast batch size and epochs, why this happens? Thanks

[rajashekar]

Hello aman, how are you??
Can you do some videos on different activation functions of neural network please??

[enumahikechukwun]

Great explanation.. Please do you have full course on deep learning in any learning platform? i will like to have your lecture in full.

[maliksalmananalyst]

Thank you so much for sharing this video. I really needed it. Sir, I want to use the firefly algorithm to optimize my CNN model's hyperparamters.
Could you please help me out with it?

[swathisree]

Sir how your fixing dense layer as 12?

[SinoFiasin]

So, we can run to find best batch and epochs then using those to find best optimizer ?

[oladelemayowa]

Hello, I keep getting two errors when I implement the GridSearchCV.

First Error is given below:
ValueError:
All the 108 fits failed.
It is very likely that your model is misconfigured.
You can try to debug the error by setting error_score='raise'.

Second Error:
TypeError: 'str' object is not callable

How do I rectify this?

[kalipza]

cant we hypertune all of them together at once, i mean finding best value of epoch, batch size, ptimizer, activation fuctions and hidden layers from the given range.. is that possible or not ?

[taiwobisade]

Hello, i tried running with a different data, but got the error below:
AttributeError: 'KerasClassifier' object has no attribute 'loss'
any idea on what could have caused this error?

[MrSoumyabrata]

What if we use training and validation data comparison to get better accuracy?

[narendrathambathka]

Bro In Real word will hyperparameter tune separately for activation functions, batch and epoch size and all??

[ajayrana]

can u explain why model cho0se sgd over adam or anyone else i mean what should the connection of sgd with data... what is going behind the wall what u think

[hanselcreado]

Dear Sir,

Please Help
After running the code

parameter_grid = dict(batch_size = batchSize, epochs=epochs)

mygrid = GridSearchCV(estimator=model, param_grid=parameter_grid, n_jobs=-1, cv=3)
grid_result = mygrid.fit(X, Y)

I am getting the error

ValueError: logits and labels must have the same shape ((None, 8) vs (None, 1))