summarization

README

Summarization

Summarize text using a GGUF language model.

How it works

1. Reads input text from --text, --file, or uses a built-in example
2. Loads a GGUF model via inference.LoadFile
3. Constructs a summarization prompt and generates a concise summary
4. Prints the summary to stdout

Usage

go build -o summarization ./examples/summarization/

# Summarize inline text
./summarization --model path/to/model.gguf --text "Your long text here..."

# Summarize a file
./summarization --model path/to/model.gguf --file article.txt

# Use the built-in example
./summarization --model path/to/model.gguf

# With GPU acceleration
./summarization --model path/to/model.gguf --device cuda

Flags

Flag	Default	Description
--model	(required)	Path to GGUF model file
--device	cpu	Compute device: "cpu", "cuda"
--text	(example)	Inline text to summarize
--file		Path to a text file to summarize
--max-tokens	256	Maximum tokens in the summary

Documentation

Overview

Command summarization demonstrates text summarization using a GGUF language model.

It reads text from a file or flag, prompts the model to produce a concise summary, and prints the result. Useful for condensing long documents, articles, or logs.

Usage:

go build -o summarization ./examples/summarization/
./summarization --model path/to/model.gguf --text "Long text to summarize..."
./summarization --model path/to/model.gguf --file article.txt