|
<< Click to Display Table of Contents >> Navigation: 3. Script Language > AI - Artificial Intelligence Commands > AIK. - Use Anthropic Claude AI |
MiniRobotLanguage (MRL)
Setting up your System for Anthropic AI
where to get a "API-Key" and more
Intention
Welcome to the Claude Family!
The dawn of AI technology has brought us to the threshold of a new era, where interaction with artificial intelligence transcends the boundaries of what we previously thought possible. At the heart of this revolution is Claude, a groundbreaking series of large language models developed by Anthropic. Designed to navigate a vast array of tasks involving language, reasoning, analysis, coding, and beyond, Claude is not just a technological marvel; it's your next step into the future of digital interaction.
1. To get started in using Claude with the SPR you will need an API-Key. Therefore you may want to first visit this Site here:
2. The next step is the same as you would do with all AI-Systems to be used with the SPR. You need to store the API-Key with the SPR.
For this execute the following Script:
' Script 1: Save the Key to the file "AIK_License_Key.dat".
' IMPORTANT: This API-Key is a phantasy API-Key it must be replaced with your API-Key.
$$KEY=sk-abcdefghijklmnopqrstuvwxyz123456
' Here we save the Keyfile at the default path, that is:
' ?exeloc\AIK_License_Key.dat
AIK.Save_Key|$$KEY
ENR.
3. The Licence-File that contains an encrypted version of your API-Key will be stored in the SPR-Folder.
And each time you want to use the AIK.-Command, it will automatically be copied to your Project Folder.
You never have to search for your API-Key and you never have an readable API-Key in your Script.
' Script 2; Using the crypted API-Keyfile
' Test if we are online, AI-Commands will only work if you are online.
NOL.
GTO.enx
EIF.
' Set Anthropic API-Key from the saved File
AIK.SetKey|from_File
' Set Model
AIK.SetModel_Completion|4
' Set Model-Temperature
AIK.Set_Temperature|0
' Set Max-Tokens (Possible lenght of answer, depending on the Model up to 2000 Tokens which is about ~6000 characters)
' The more Tokens you use the more you need to pay.
AIK.SetMax_Token|25
' Ask Question and receive answer to $$RET
AIK.Ask_Claude|Wieviel Uhr ist es?|$$RET
DBP.$$RET
:enx
ENR.
Introduction: This manual provides a comprehensive guide to understanding and reducing latency in your Claude-powered applications. It covers various terms, measurements, and strategies to help you improve the speed and performance of your projects.
Section 1: Understanding Latency
1.1 Baseline Latency: This refers to the time taken by the model to process the prompt and generate a response, without considering the input and output tokens per second. It gives a general idea of the model's speed.
1.2 Time to First Token (TTFT): This metric measures the time it takes for the model to generate the first token of the response, from when the prompt was sent. It's particularly relevant when using streaming and aiming to provide a responsive user experience.
Section 2: Strategies for Reducing Latency
2.1 Choose the Right Model: Selecting the appropriate model for your use case is a straightforward way to reduce latency. Anthropic offers three models:
•Claude 3 Haiku: The fastest model, ideal for applications requiring quick responses and tolerating a smaller model size.
•Claude 3 Sonnet: Balances speed and model size, offering better performance than Haiku while maintaining relatively fast latency.
•Claude 3 Opus: The largest and most powerful model, perfect for complex tasks demanding high-quality output, but may have higher latency.
2.2 Optimize Prompt and Output Length: Reducing the number of tokens in your input prompt and expected output can decrease latency. Here are some tips:
•Be clear and concise in your prompt.
•Ask Claude to be concise in its responses.
•Set appropriate output limits using the AIK.Set Max Token Command.
•Experiment with the temperature parameter to control output randomness.
Section 3: Wrapping Up
Reducing latency is crucial for building responsive applications. By choosing the right model, optimizing your prompts and outputs, you can significantly improve the speed and performance of your Claude-powered projects.
See also: