ai_app_ollama 1.0.1

A Dart Plugin which can be used to make api request to Ollama AI on localhost

Ollama ai application library for dart and flutter

uses:

Get up and running with large language models locally.

Linux:

curl https://ollama.ai/install.sh | sh

Quickstart:

To run and chat with Llama 2:

ollama run llama2

Model library Ollama supports a list of open-source models available on ollama.ai/library

Here are some example open-source models that can be downloaded:

Model Parameters Size Download Mistral 7B 4.1GB ollama run mistral Llama 2 7B 3.8GB ollama run llama2

If you want to generate response from a model, you can use the ask method. This method takes a prompt and a model name, and returns a CompletionChunk object.

void main() async { // Create an Ollama instance final ollama = Ollama();

// Generate a response from a model final response = await ollama.ask('Tell me about llamas', model: 'llama2');

// Print the response print(response.text); }

Stream responses can be generated using the generate method. This method takes a prompt and a model name, and returns a Stream

void main() async { // Create an Ollama instance final ollama = Ollama();

// Generate a response from a model final response = ollama.generate('Tell me about llamas', model: 'llama2');

// Print the response await for (final chunk in response) { stdout.write(chunk.text); } }