Machine Learning / AI

Who to show in machine marketing?

Why is dropout disabled during inference?

Is there confidence in the classification tree?

Are you familiar with RTB auctions?

In which language is it better to implement model training?

What is LoRA merging and how to merge multiple adapters?

Middle — Senior

HybridHeadHunter

Name the loss function for logistic regression.

What is sliding-window attention in Mistral?

Middle — Senior

HybridСовкомбанкМТС AI

What stop criteria for tree growth exist besides depth limitation?

What type of computation graph is used in PyTorch — dynamic or static?

Middle

HybridИннотех

How does WordPiece differ from BPE in terms of pair selection for merging?

How to build a sample for training a model with relevant observation pairs?

What is the concept of planner-executor separation in agents?

How to classify issuing a new or old identifier?

How do you obtain binnings for new observations?

How is dropout implemented in PyTorch?

How to avoid loss of fidelity when mixing features?

How does a tree return the probability of an object belonging to a class?

What internal criteria determine the terminality of a tree leaf?

What is the difference between log loss and cross-entropy?