Stanley LarryStanley

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

You are an expert in prompt engineering, specializing in optimizing AI code assistant instructions. Your task is to analyze and improve the instructions for Claude Code. Follow these steps carefully:

Analysis Phase: Review the chat history in your context window.

Then, examine the current Claude instructions, commands and config <claude_instructions> /CLAUDE.md /.claude/commands/*

通览了 casbin 的文档，结合先前对 AWS IAM 的理解，以及对 ladon SDK 的使用，总结对比一下 Ladon & Casbin 两个授权库。

1. 项目定位

先对比两个项目的简介：

ladon

A SDK for access control policies: authorization for the microservice and IoT age. Inspired by AWS IAM policies. Written for Go.

How to train your own object detection models using the TensorFlow Object Detection API (2020 Update)

This started as a summary of this nice tutorial, but has since then become its own thing.

Prerequisites
Annotating images and serializing the dataset
Choosing a neural network and preparing the training pipeline
Training the network

What you need:

Raspberry Pi RevB
Raspberry Pi NoIR camera

Download raspbian and install to SD-card. Here is a good guide on how to install the .img file to the card: https://www.andrewmunsell.com/blog/getting-started-raspberry-pi-install-raspbian

Connect ethernet or wifi, then boot the Raspberry Pi and follow configuration options. Make sure you enable the camera option.

When you get to the command-line:

	#!/usr/bin/python

	import random
	import numpy as np
	from matplotlib import pyplot as plt

	"""
	Straight-forward implementation of PO2C.
	"""
	def po2c(num_buckets, num_groups, group_size):

	import tensorflow as tf
	import random
	import numpy as np
	import time

	BASIC_HOME = "/Users/allwefantasy/Downloads"
	WORD_VECTOR_FILE = BASIC_HOME + '/laiwen/zhuhl_laiwen_word_embedding'
	WORD_FILE = BASIC_HOME + '/laiwen/zhuhl_laiwen_keywords2'
	WORD_RESULT_VECTOR_FILE = BASIC_HOME + '/laiwen/WORD_RESULT_VECTOR_FILE4'
	MODEL_SAVE_DIR = BASIC_HOME + '/laiwen/model/autoencoder'

	#Source code with the blog post at http://monik.in/a-noobs-guide-to-implementing-rnn-lstm-using-tensorflow/
	import numpy as np
	import random
	from random import shuffle
	import tensorflow as tf

	# from tensorflow.models.rnn import rnn_cell
	# from tensorflow.models.rnn import rnn

	NUM_EXAMPLES = 10000

	# docker-neuraltalk2 demo with webcam support (Warning: docker-neuraltalk2 image size > 3GB)

	# Launch interactive container:
	docker run -i -t -p 8000:8000 --privileged -v /dev/video0:/dev/video0 beevaenriqueotero/docker-neuraltalk2 /bin/bash
	# And then execute:
	git clone https://github.com/Mithul/neuraltalk2.git neuraltalk2-camera
	luarocks install camera
	cd neuraltalk2-camera/vis
	python -m SimpleHTTPServer &
	cd ..

	// the main app file
	import express from "express";
	import loadDb from "./loadDb"; // dummy middleware to load db (sets request.db)
	import authenticate from "./authentication"; // middleware for doing authentication
	import permit from "./authorization"; // middleware for checking if user's role is permitted to make request

	const app = express(),
	api = express.Router();

	// first middleware will setup db connection