Neural Network Challenge-: House price prediction

#81

hey @sambher_shubham , i am might be lacking in correcting the directories .
can you please provide me 2 or 3 names of images in covid directory.

#82

have a look at this link
https://colab.research.google.com/drive/1-ZBCnEr5yrhcUiRK_kEZ-vks5AdBZDG5?usp=sharing

Run it once and let me know , did i have made the directories properly.

#83

Yes bhaiya you have made the directories properly. I got my mistake…

#84

what was your mistake ?

#85

While extracting images from kaggle dataset i was extracting images from test folder but the images should be extracted from the train folder.

#86

Thank You Soo Much Bhaiya :grin:

#87

ohhh…lol.
No problem. At least you got your error. That’s what really matters.

#88

Any more doubts can i ask you here in this thread link?

#89

it would be good if you raise them in another doubt thread. :sweat_smile:.

#90

Koi aur toh ni aa jaega dusre doubt me :rofl: :rofl:

#91

nhi nhi…
main hi doubt lunga. Dont worry.

#92

On an unseen data this model is giving between 85 to 90 % accuracy.

#93

Like in kaggle dataset there is a folder named as test. For that folder it is not predicting as good results as on validation set

#94

It will never perform the same as it does on validation there is always some difference.
Getting a 90% score is great , although if still you want to improve your score , then you can try different other experiments on the current model and check how does it perform on the test set.
If it improves then update your model else try something new.
You can try , Using different Checkpoints , playing with learning rate , optimizers , layers etc.

#95

Okay.I will try changing the parameters.

#96

Its giving 80 percent on the test normal images :sweat_smile: :sweat_smile:

#97

Yeah it will get effected a bit , so just try with more different experiments.
There will be some , which improve your score.

#98

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.