Pavel Zloi EvilFreelancer

Improving Gemini's Text Generation Accuracy with Corpus Managed by Google Spreadsheet as RAG

Abstract

Gemini excels at text generation with RAG for large datasets, but smaller ones benefit from prompting or data upload. This report explores using Gemini 1.5 Flash/Pro with RAG on medium-sized, Google Spreadsheet-stored datasets for improved accuracy and effectiveness.

Introduction

Deploy K3s and Rancher on Raspberry Pi 4 cluster

Today I tried to setup a small Kubernetes cluster on top of 3 Raspberry Pi 4 (4GB Memory). Here is the steps to install the cluster.

Preparation

I have 3 Raspberry Pi 4 stacked with PoE headers and connected to a PoE switch at home. The are connected to Internet through a home router. All Pis are equipped with a 64GB Samsung SDXC card flushed with Ubuntu 20.04 image.

This was tested on a ThinkPad P70 laptop with an Intel integrated graphics and an NVIDIA GPU:

lspci | egrep 'VGA|3D'
00:02.0 VGA compatible controller: Intel Corporation Device 191b (rev 06)
01:00.0 VGA compatible controller: NVIDIA Corporation GM204GLM [Quadro M3000M] (rev a1)

A reason to use the integrated graphics for display is if installing the NVIDIA drivers causes the display to stop working properly. In my case, Ubuntu would get stuck in a login loop after installing the NVIDIA drivers. This happened regardless if I installed the drivers from the "Additional Drivers" tab in "System Settings" or the ppa:graphics-drivers/ppa in the command-line.

	from transformers import (
	AutoConfig,
	AutoTokenizer,
	BitsAndBytesConfig,
	GenerationConfig,
	AutoModelForCausalLM,
	LlamaTokenizerFast,
	PreTrainedModel,
	TextIteratorStreamer,
	StoppingCriteria,

	import groovy.json.JsonSlurper
	import groovy.json.JsonParserType
	import org.sonatype.nexus.repository.storage.Asset
	import org.sonatype.nexus.repository.storage.Query
	import org.sonatype.nexus.repository.storage.StorageFacet
	import java.util.concurrent.ExecutorService
	import java.util.concurrent.Executors

	CONDA_CLOUD_BASE_URL = "https://conda.anaconda.org"
	NEXUS_REPOSITORY = "http://127.0.0.1:8081/repository/conda"

	############ REPLACE app/models/enterprise_token.rb in the source code with this file! ################
	############ also be sure to RESTART OpenProject after replacing the file. ################
	############ it doesn't show that enterprise mode is enabled in the settings, but all ################
	############ enterprise mode features, such as KanBan boards, are enabled. ################
	#-- copyright
	# OpenProject is an open source project management software.
	# Copyright (C) 2012-2023 the OpenProject GmbH
	#
	# This program is free software; you can redistribute it and/or
	# modify it under the terms of the GNU General Public License version 3.

	###################################################
	##
	## Alertmanager YAML configuration for routing.
	##
	## Will route alerts with a code_owner label to the slack-code-owners receiver
	## configured above, but will continue processing them to send to both a
	## central Slack channel (slack-monitoring) and PagerDuty receivers
	## (pd-warning and pd-critical)
	##

	<?php

	namespace App\Models;

	use Illuminate\Database\Eloquent\Model;

	class City extends Model
	{
	/**
	* @var array

	const axios = require("axios");

	const urls = ['https://github.com/mzabriskie/axios', 'https://discord.js.org'];

	async function asyncLoadAllTheChunks() {
	const resps = await Promise.all(urls.map(url=> axios.get(url)))
	const [first, second] = resps;
	console.log(`
	тут первый ${first}
	тут второй ${second}

	// ...
	"scripts": {
	"build": "neo build --config webpack.prod.js",
	"start": "neo start --port 3001 --config webpack.dev.js",
	// ...
	}
	// ...