What is Residual Connection (residual links)?
Machine Learning / AI
Add two numbers in one function, with a decorator that first prints 'Hello', then calls the addition function, and then prints 'Bye'. def hello(func): def wrapper(a,b): print('Hello') result = func(a,b) print('Bye') return result return wrapper @hello def add_numbers(a,b): return a+b print(add_numbers(2,3))
What is ReLU and why is it used?
Tell us about your experience.
Tell about gradient descent: what types are there?
Do you know anything about mutable and immutable data types, and which data belongs to which?
Please tell how dictionaries work under the hood in Python, and what about collisions?
Is it possible to use a decorator without the @ syntax?
Pre-Layer Norm vs Post-Layer Norm — what is the difference and why is it important for training stability?
Why is non-linearity needed in neural networks?
So, the tool didn't work with the model — it only accessed the knowledge base?
In this scheme, if a user writes the word "hello", it will go into RAG and get a response through RAG for the word "hello". How can we see/prevent this?
Can we use the same metrics when measuring quality by top-K and the metric asked from the user (like/dislike)? Or are these different things?
How to deal with the problem of increasing context when actively engaging in a dialogue with a user?