Satyam Mishra

Could you help me understand more about your background by giving a brief introduction of this booking? My name is Atir Mishra, and, uh, I am a data scientist. Um, I worked with Radcliffe Labs And where I worked on, uh, building, uh, where I worked on building recommendation system. And, um, at at Amaruzala, I worked on also worked on a recommendation system, and I Develop the recommendation system that, uh, caters to them. Actually, millions of users. And, At, uh, Radcliffe Labs, I designed computer vision. There's a problem, uh, task for fluoro grooming, Uh, detection problem. And, uh, so this is a brief introduction about myself. And I did BCA from Igno and, uh, I had some certification from, um, IIT Madras from tactical machine learning and TensorFlow. And I, uh, also have a certification from Datachem for, Uh, power data scientist strike with Python. So And, um, this is it.

Which, uh, I think it has to be most efficient for indexing when working on machine learning problem.

What approach, uh, actually, what approach would I take to handle imbalanced classes in dataset when training a deep learning model using PyTorch. I can, uh, basically, apply techniques like ensemble methods. I can also apply resampling techniques. I can also apply data augmentation techniques. And I also can also apply weighted losses, technique, which is like, uh, which is a pi PyTorch has inbuilt functionality of a cross entropy loss. And, also, it has some other functionality like that. Precision recall and f one score to evaluate

Use case for Python, Lambda function in preprocessing data for machine learning model. One of the best use case I found in, Using Lambda function is for Python and data is that I I can just build a function, Sorted from version of function, uh, like, uh, that's required for cleaning the up up and start the specific data. Like, uh, I wanted to do do some scaling, Or maybe I can also do some other things like,

Can you mention a Python feature that helps prevent over putting during model training? I don't know. Can you mention a pie talk feature that help prevent over fitting during model training? Uh, I don't know that, actually.

Uh, what, uh, I would use, Recursive feature elimination or, Lasso regression for features selection for large scale machine learning project.

The one below is a code in r. It is supposed to give up the mean of a numeric vector. Can you name debug whites? Do you be giving The necessary. Why it's giving wrong output and suggest the necessary correction? Uh, let me check. Use mean calculation function, some value. Mean conclusion. Actually, this code has a lot of issues. Uh, not not not a lot of issues, but Basically, 2 issue. The 1 issue is that, um, that the, uh, the inner loop the loop of that is actually It's calculating the sum in each iteration, but they're not updating some value correctly. And, uh, the second thing is that I need to we need to move the initialization of some value outside the loop and fix that. Actually, that is all correct. Maybe just we just move the any slicing of some of the loop outside the loop only.

Uh, in the neural network, training loop written in the Python, identify the section of code. Uh, in neural network training loop written in that Python, Uh, identify that section of the loop code that might cause an error during back propagation. What is the and that import while touching what quite might cause in that provision and explain why. Uh, so it really is what you need to do. This might be problematic. Mod.eval. Okay. Uh, so how I should I do that? And then not relu. Uh, and are not linear. I've crossed 2, uh, long dot tensor. The line target dot on this typo, it should be long tensor 1. That's

Which when hyper tuned are parameters for PyTorch model, which method do you prefer? Good search or random search? Hyperparameter tuning. Uh, I would use grid search.

When developing a PyTorch model, how do you ensure efficient memory usage during training, especially with a large How do you ensure effective, efficient memory using training, Especially with the large dataset. Okay. When developing a pilot class model, how do you Ensure efficient memory is during training, especially with the larger So, uh, in, uh, so When developing a PyTorch model to ensure the efficient memory usage, In PyTorch, I would consider using techniques like batch loading, data augmentation, optimizing your our model architecture for memory. Additionally, monitored GPU memory usage and clearing unnecessary variable can also have help memory effectively.