Meta released its official Prompt Engineering Guide (Tutorial included)

More A.I. news from Igor & the Skill Leap AI Team:

Plus New Meta Promptin guide ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

ChatGPT has a brand new update that makes it possible to easily access GPTs within any conversation. This update will give ChatGPT abilities we've never had before.

Midjourney has also released a major update that will change how we all use Midjourney to achieve consistent results.

AI Headlines at a Glance:

New ChatGPT Update - Bring GPTs into any conversation
Meta released its official Prompt Engineering Guide (Tutorial included)
Midjourney introduces “Consistent Styles”
New Tutorial: Create an Epic Movie with AI - Step-by-Step Guide

Top AI News of the Week

New ChatGPT Update - Bring GPTs into Any Conversation

ChatGPT just introduced one of the most useful updates, which is going to change how we use CustomGPTs.

You can now bring GPTs into any conversation in ChatGPT - simply type @ and select the GPT.

This allows you to add relevant GPTs with the full context of the conversation. This works for public, unlisted, and private GPTs.

Here is a complete tutorial to showcase the power of GPTs in the same chat with practical examples.

Meta's Official Prompt Engineering Guide

Meta is behind one of the most popular large language models called Llama 2, and they have now released a prompt engineering guide that we can use to get better results when using any large language model, including ChatGPT.

This document is created for developers, so we simplified it for non-developers and created a tutorial highlighting seven fundamental prompting techniques that will take your prompting to the next level.

1 - Explicit Instructions

Detailed, explicit instructions produce better results than open-ended prompts.

2 - Zero-Shot Prompting

Large language models like Llama 2 are unique because they are capable of following instructions and producing responses without having previously seen an example of a task. Prompting without examples is called "zero-shot prompting."

3 - Few-Shot Prompting

Adding specific examples of your desired output generally results in more accurate, consistent output. This technique is called "few-shot prompting."

4 - Role Prompting

Large Language Models often give more consistent responses when given a role, such as “Act as a Digital Marketer.”

5 - Chain-of-Thought

Simply adding a phrase encouraging step-by-step thinking significantly improves the ability of large language models to perform complex reasoning.

6 - Self-Consistency

Self-Consistency introduces enhanced accuracy by selecting the most frequent answer from multiple generations.

7 - Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) is a technique where AI enhances its responses by first gathering information from a large database of texts and then using that information to answer questions or generate content. Here is a direct link to the guide: https://github.com/facebookresearch/llama-recipes/blob/main/examples/Prompt_Engineering_with_Llama_2.ipynb

We also created a tutorial that includes example prompts for each prompting technique.

Midjourney introduces “Consistent Styles'”

Midjourney rolled out a huge update to create images in a consistent style by using a reference photo. This new update is called “Style Reference” and it's designed with a new algorithm from the ground up.

Even though a similar option was available before, this new update works much better with a lot more control over the output.

How to use Style References

Type --sref after your prompt and put one (or more) URLs to images like this --sref urlA urlB urlC
The image model will look at the image urls as 'style references' and try to make something that 'matches' their aesthetics
Set relative weights of styles like this --sref urlA::2 urlB::3 urlC::5
Set the total strength of the stylization via --sw 100 (100 is default, 0 is off, 1000 is maximum)
Regular image prompts must go before --sref like this/imagine cat ninja ImagePrompt1 ImagePrompt2 --sref stylePrompt1 stylePrompt2
This works for both V6 and Niji V6 (it does not work with V5 etc)

AI Tutorial of the week

Create an Epic Movie with AI - Step by Step Tutorial: https://www.youtube.com/watch?v=QiUY9HpHbHM

Sent by Howfinity LLC. 1500 N. Halsted St. • Chicago, IL • 60642

0 comments