Name the two practical limits mentioned in the source that prevent indefinite scaling of model size and data.
Question: In one to three sentences, name the two practical limits described in the source that prevent indefinitely scaling neural network size and training data.
Sample answer: The two practical limits are that training very large models is slow, causing computational problems, and that the supply of additional training data can be exhausted.
Key points:
- Computational problems from slow training of large models
- Exhaustion of available additional training data
Rubric: Full credit requires mentioning both the computational/slow-training limit and the data exhaustion limit.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Which practical constraint eventually limits the strategy of continually increasing neural network size?
True or False: In principle, unlimited increases in network size and training data can perform well on many learning problems.
Training very large models is _____, which creates a practical limit on scaling model size.
Match each scaling strategy to the practical limit it eventually encounters.
Order the reasoning steps for why scaling model size and data has practical limits.
Explain why the theoretical benefit of unlimited scaling does not hold in real-world machine learning practice.
A team wants to keep scaling their model indefinitely to fix bias and variance issues. What should they consider?
Name the two practical limits mentioned in the source that prevent indefinite scaling of model size and data.
What limits a team's ability to keep increasing training data as a strategy?
True or False: Training very large neural networks is described as fast and free of computational problems.