Share Your Work Here - Assignment 2

Hi all, my work on assignment2, working on optional segment now hope to finish and write a blog post:

10 Likes

Wow, that makes a lot of sense now. Thank you!

1 Like

For Lecture 1, I used colab. This time I am using kaggle. I get this error for the first run. can u tell me what is wrong?

!pip install jovian --upgrade --quiet
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by ‘NewConnectionError(’<pip._vendor.urllib3.connection.VerifiedHTTPSConnection object at 0x7f3bdb72fa50>: Failed to establish a new connection: [Errno -3] Temporary failure in name resolution’)’: /simple/jovian/

i think you have to enable internet, go to the right side menu for the settings and enable it.

1 Like

doesnt seem to work. i verified the internet for phone number after receiving the sms. That verify button didnt give me any output.

The colab works fine.

22 posts were merged into an existing topic: Share your work : June 1 - June 7

Hello everyone,
Here is the my second assignment link: Insurance linear regression

3 Likes

Greetings folks,
Here’s my working notebook : mynotebook .

2 Likes

Hi everyone,
I completed the second assignment, available here: https://jovian.ml/benrallet/02-insurance-linear-regression
Happy to get some feedback or any suggestion how to improve my model
Thanks,
Benedicte

1 Like

Here is my notebook. Please let me know if you see some scope for improvement

1 Like

Hello everyone, here’s the notebook for second assignment : Second Assignment

1 Like

Hey guys, here’s my notebook. Lemme know if there’s anything more I could do to improve.

Insurace Linear Regression
The above one uses l2 norm.

Here’s an improvement on the Model thanks to @jae0143, it now has better results
New Linear Regression Model for Insurance Assignement (02)

2 Likes

Hey guys,

My Insurance Linear Regression notebook link. If you guys get a significantly better RMSE score or R-squared, please let me know what you did differently.

My RMS Loss : 23k (validation loss)
R-squared Value : 0.74

2 Likes

Hi fellow ML learners,

Here’s my submission for Assignment #2: Train your first model.

Details of the model are as follows:
Train size, value size: (1043, 228) 0.18
Batch size: 4
Loss function: l1_loss
Epoch, Learning rate: 1000 with variable learning rate (i.e. 1e-2, 1e-3, 1e-4, 1e-5, 1e-6)
Final value loss is ~7100.

Feedback on how to further improve my model is appreciated. Thank you! :slightly_smiling_face:

1 Like

Hi everyone!

This is my second assignment jupyter notebook.

1 Like

Hey, here is my submission for the second assignment.

Split rate: 0.15
Loss function: Mean Absolute Error
My loss is around 7500

Some predictions from the model:
Target: 14735.5098 - Prediction: 14239.6641
Target: 13196.7881 - Prediction: 12573.8662
Target: 4255.1924 - Prediction: 6493.3584

I will add the blogpost when I am done with it.

2 Likes

As seen above, using l1_loss instead of mse_loss as your loss function results in much better accuracy (lower val_loss). Give that a try!

1 Like

Hello everyone,
I completed the second assignment based on Insurance.
My Assignment notebook is available here:

Assignment-2-Insurance-Linear-Regression

Feel free to give some suggestions or feedback.

3 Likes

Thank you for your feedback. I was able to bring down the val_loss to 4865.0591.

2 Likes

I used Huber loss(Smooth mean absolute error) in notebook, it is F.smooth_l1_loss. It showed a great result!

1 Like