Chapter
The Accessibility of Deep Learning
52:03 - 56:27 (04:23)
Lex Fridman discusses the accessibility of deep learning and the use of multi-GPU or multiple machine training to speed up code. Many people believe that deep learning is inaccessible to individuals outside of Google for useful work.
Clips
The speaker shares their experience with distributed image recognition on ImageNet and why they think multi-GPU and multiple machine training is largely a waste of time.
52:03 - 53:05 (01:01)
Summary
The speaker shares their experience with distributed image recognition on ImageNet and why they think multi-GPU and multiple machine training is largely a waste of time.
ChapterThe Accessibility of Deep Learning
EpisodeJeremy Howard: fast.ai Deep Learning Courses and Research
PodcastLex Fridman Podcast
Some researchers believe that relying on big data sets can stifle creativity in AI development, and smaller data sets can yield just as useful results.
53:05 - 54:45 (01:40)
Summary
Some researchers believe that relying on big data sets can stifle creativity in AI development, and smaller data sets can yield just as useful results. This has led to a reluctance to pursue deep learning due to perceived inaccessibility to those outside of large tech companies like Google.
ChapterThe Accessibility of Deep Learning
EpisodeJeremy Howard: fast.ai Deep Learning Courses and Research
PodcastLex Fridman Podcast
De-Oldify is a tool that utilizes the power of machine learning to add color to black and white photos and films, without needing multiple GPUs.
54:45 - 56:27 (01:41)
Summary
De-Oldify is a tool that utilizes the power of machine learning to add color to black and white photos and films, without needing multiple GPUs.