Read on Twitter
GPT-3: one of the world's most advanced machine learning models.
Let's talk about it.
(will it take your job?)
Let's talk about it.
(will it take your job?)
Generative Pre-trained Transformer 3 or GPT-3 for short, is an autoregressive language model that uses deep learning to produce human-like text created by this company called OpenAI.
Sounds confusing? Let's break it down.
(2 / 6)
Sounds confusing? Let's break it down.
(2 / 6)
In plain English, GPT-3 is a machine learning model that predicts a future word from a set of words given a context.
It like I tell say "Tesla's CEO?",you'd reply "Elon Musk", this is sort of what GPT-3 does.
(3 / 6)
It like I tell say "Tesla's CEO?",you'd reply "Elon Musk", this is sort of what GPT-3 does.
(3 / 6)
The data that was used to train GPT-3 was mostly from wikipedia and guess what, wikipedia is HUGE.
Paragraphs were taken from these wikipedia articles and randomly, words would be removed. It was GPT-3 task's to determine which word fit in there.
(4 / 6)
Paragraphs were taken from these wikipedia articles and randomly, words would be removed. It was GPT-3 task's to determine which word fit in there.
(4 / 6)
It cost OpenAI nearly $4.6 million to train GPT-3, this is because it has about 175 billion parameters compared to the 1.5 billion parameters GPT-2 had, a gigantic leap.
You can use GPT-3 for things like:
- Generating code
- Designing applications
- Explaining code
(5 / 6)
You can use GPT-3 for things like:
- Generating code
- Designing applications
- Explaining code
(5 / 6)
Will GPT-3 take your job?
No, at least not right now, it is still not perfect but this is something to be genuinely worried about.
What do you think about GPT-3?
(6 / 6)
No, at least not right now, it is still not perfect but this is something to be genuinely worried about.
What do you think about GPT-3?
(6 / 6)
