How to evaluate the quality of a retriever? Where to get a collection of documents for metric evaluation?
Senior
380
Machine Learning / AI
3611
0
What is vLLM and what is it used for?
Senior
191
What is the difference between Tensor Parallel Size and Data Parallel Size in vLLM?
Senior
185
You have a small dataset with movie review texts in Russian and their binary sentiment labels (labels: 0 — negative, 1 — positive). Your task is to implement a complete pipeline for training a simple text classification model within 15 minutes without using pre-built encoders like BERT, and to output the accuracy on this data. reviews = [ "This movie is simply amazing!", "Horrible acting and boring plot.", "I enjoyed every moment!", "The worst movie I have ever seen.", "Great cinematography and a wonderful story.", "Unbearably boring and predictable." ] labels = [1, 0, 1, 0, 1, 0]
Senior
183
What is RAG and what fundamental problem does it solve?
Senior
181
Which parameter in the embedding model determines the maximum number of tokens at the input? What happens if more tokens than the limit are provided?
Senior
168
How does increasing the batch size fourfold affect the learning rate and why? How does backpropagation work?
Senior
165
Tell about yourself and your experience.
Senior
162
What is an embedding model and what is a token?
Senior
155
What is model quantization and what happens to the weights when it is applied in vLLM?
Senior
155
How are data extracted from a vector database based on a user query?
Senior
153
What happens if you provide a context of 120,000 tokens to an LLM with a limit of 60,000?
Senior
151
How was the information search organized in the vector database? Why was Qdrant used?
Senior
141
What is LangGraph, what is it used for, how are branches and conditional transitions implemented? How do tools differ from nodes?
Senior
140