3 2 weeks ago

Raspberry Pi 5 Optimized Sentiment Analysis Checkpoint

ollama run jewelzufo/granite-4.0-h-350m-rpi5-sentiment-analysis

Details

2 weeks ago

726de3d7ea41 · 223MB ·

granitehybrid
·
340M
·
Q4_K_M
{{ if .System }}<|start_of_role|>system<|end_of_role|> {{ .System }}<|end_of_text|> {{ end }}{{ if .
TASK: CLASSIFICATION FORMAT: JSON_ONLY SCHEMA: { "sentiment": "POSITIVE" | "NEGATIVE" | "NEUTRAL", "
{ "num_ctx": 2048, "num_thread": 4, "repeat_penalty": 1.2, "stop": [ "<|end_

Readme

Overview

This Modelfile checkpoint is customized for local use on a Raspberry Pi 5 (8gb-16gb) for basic sentiment analysis tasks.

Raspberry Pi 5 Considerations

Cooling: Active cooling is highly recommended to avoid CPU throttling issues.

Quick Start

When chatting with the model, simply enter the text (short) to be classified. It will output a json format sentiment.