Jason Ross rossja

Official

Each of these commands will run an ad hoc http static server in your current (or specified) directory, available at http://localhost:8000. Use this power wisely.

Discussion on reddit.

Python 2.x

$ python -m SimpleHTTPServer 8000

HuggingFace SF_Convertbot Scanner

This script is designed to assist in identifying pull requests to HuggingFace repositories that are sourced from the SFConvertbot user.

The SFConvertbot user is part of an automated tool used by HuggingFace to provide safetensor versions of models. As published by HiddenLayer this bot can be used by malicious actors to potentially insert malicious content into models.

This tool is a simple script to query all models released by a HuggingFace author, and checks all

ANCIENT
ANGEL
ANGULAR
ARMADILLO
ARMOR
ASH
Backpack
BAD DOG
BAIT
BAT

RAG Implementation with Google Gemini and ChromaDB

A simple example implementation of Retrieval Augmented Generation (RAG) using Google's Gemini API and ChromaDB for document storage and retrieval.

Overview

This code demonstrates a basic RAG system with three main components:

Document embedding using Google's text-embedding-004 model
Vector storage and retrieval using ChromaDB
Question answering using Gemini 1.5 Flash model

	/**
	* utils module - provides various utility functions used by the application
	* @module utils
	*/

	/**
	* A function to sort an array of indices by a given property
	* put simply: this lets you sort an array of json objects by
	* whatever property you tell it to use.
	* Example: `var teams = data.teams.sort(utils.GetSortOrder("name"))`

	"""
	presumes a list of slack tokens (1 per line)
	in a file at ./tokens.txt
	runs through them one by one using the slack auth.test endpoint
	returns info if they are valid, otherwise no
	"""
	import requests

	url = 'https://slack.com/api/auth.test'

	import torch

	if torch.cuda.is_available():
	print ("cuda available")
	print(f'available devices: {torch.cuda.device_count()}')
	elif torch.backends.mps.is_available():
	print ("mps available")
	device = torch.device("mps")
	x = torch.ones(1, device=device)
	print (x)

	# smuggle text hidden as invisible unicode chars
	# credit to jthacker: https://x.com/rez0__/status/1745545813512663203
	# and embrace the red: https://embracethered.com/blog/posts/2024/hiding-and-finding-text-with-unicode-tags/

	import pyperclip

	def convert_to_tag_chars(input_string):
	return ''.join(chr(0xE0000 + ord(ch)) for ch in input_string)

	# Example usage:

	<html>
	<!--
	// neat font for syntax highlighting. see https://blog.glyphdrawing.club/font-with-built-in-syntax-highlighting/
	// font download at https://blog.glyphdrawing.club/assets/fonts/FontWithASyntaxHighlighter-Regular.woff2
	-->
	<head>
	<style type="text/css">
	@font-face {
	font-family: 'FontWithASyntaxHighlighter';
	src: