Categorygithub.com/ztkent/ai-util
modulepackage
0.7.2
Repository: https://github.com/ztkent/ai-util.git
Documentation: pkg.go.dev
Overview
Versions
1
Dependencies
18
Dependents
5
Files
835 SLOC

# README

AI Util

A unified platform to build apps with AI.

Features

  • Supported AI Providers:
  • Conversation Controls
  • Token Management + Limits
  • Resource Management

Installation

go get github.com/ztkent/ai-util

Example

    client, _ := aiutil.NewAIClient("openai", "gpt-3.5-turbo", 0.5)
    conversation := aiutil.NewConversation("You are an example assistant.", 100000, true)
    response, _ := client.SendCompletionRequest(CtxWithTimeout, conversation, "Say hello!")

Required API Keys

ServiceEnvironment Variable
OpenAIOPENAI_API_KEY
ReplicateREPLICATE_API_TOKEN

Available Models

OpenAI Models

Model NameModel IdentifierCost (IN/OUT per 1M tokens)
GPT-3.5 Turbogpt-3.5-turbo$0.50 / $1.50
GPT-4gpt-4$30.00 / $60.00
GPT-4 Turbogpt-4-turbo$10.00 / $30.00

Replicate Models

Model NameModel IdentifierCost (IN/OUT per 1M tokens)
Meta Llama 2-70bmeta/llama-2-70b$0.65 / $2.75
Meta Llama 2-13bmeta/llama-2-13b$0.10 / $0.50
Meta Llama 2-7bmeta/llama-2-7b$0.05 / $0.25
Meta Llama 2-13b Chatmeta/llama-2-13b-chat$0.10 / $0.50
Meta Llama 2-70b Chatmeta/llama-2-70b-chat$0.65 / $2.75
Meta Llama 2-7b Chatmeta/llama-2-7b-chat$0.05 / $0.25
Meta Llama 3-8bmeta/meta-llama-3-8b$0.05 / $0.25
Meta Llama 3-70bmeta/meta-llama-3-70b$0.65 / $2.75
Meta Llama 3-8b Instructmeta/meta-llama-3-8b-instruct$0.05 / $0.25
Meta Llama 3-70b Instructmeta/meta-llama-3-70b-instruct$0.65 / $2.75
Mistral 7Bmistralai/mistral-7b-v0.1$0.05 / $0.25
Mistral 7B Instructmistralai/mistral-7b-instruct-v0.2$0.05 / $0.25
Mixtral 8x7B Instructmistralai/mixtral-8x7b-instruct-v0.1$0.30 / $1.00

# Functions

No description provided by the author
Generate a resource message based on the path and type, return the message to append to the conversation.
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
Estimate the number of tokens in the message using the OpenAI tokenizer.
No description provided by the author
No description provided by the author
Ensure we have the right env variables set for the given source.
Determine if the user's input contains a resource command There is usually some limit to the number of tokens.
No description provided by the author
Start a new conversation with the system prompt A system prompt defines the initial context of the conversation This includes the persona of the bot and any information that you want to provide to the model.

# Constants

No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
OpenAI Models.
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
Open-Source Models via Replicate.
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author
No description provided by the author

# Structs

No description provided by the author
The OAIClient struct is a wrapper around the OpenAI client.
No description provided by the author

# Interfaces

No description provided by the author

# Type aliases

No description provided by the author
No description provided by the author
No description provided by the author