Sandro Paganotti's Blog

LinkedIn has become the go-to platform for professionals, businesses, and entrepreneurs looking to build a credible online presence. Having a high follower count is essential for networking, personal branding, and establishing authority in any industry. However, growing an audience organically can be slow and challenging, leading many professionals to buy LinkedIn followers to speed up the process.

Choosing the right provider is crucial to ensuring that you receive real, high-quality follower...

Following from my previous post I’ve explored with an algorithm to create Sudokus by iterating from an empty grid by adding valid numbers to random cells and then invoking the solving function twice, once navigating the solution space from 1 to 9 and the other in reverse. If both solving functions return the same, valid, grid then that’s the solution.

from pprint import pprintfrom math import floorfrom copy import deepcopyfrom itertools import chain from random import shuffledef solve(sr...

Just a small practice exercise that solves Sudoku by iterating over valid values for each cell and backtracking on errors. I like how compact the solution is, although probably not very performant.

from pprint import pprintfrom math import floorfrom copy import deepcopygrid = []content = Nonewith open('startgrid.txt', 'r') as file: content = file.read().replace("\n", " ").split(" ")for row in range(9): grid.append([]) for col in range(9): grid[row].append(int(content[row*9...

FAISS implements similarity search. Sentence Transformers encodes sentences into vectors FAISS can use. Llama can be configured to perform semantic searches in a FAISS vector store. By combining all of the above together we can build a local GenAI powered assistant capable of performing semantic queries to extract information from a local corpus of documents, let’s see how.

For this example I’m using llama-2-7b-chat.ggmlv3.q4_0.bin and this document (notes.txt) I’d like my agent to provide me a...

I’ve made a digital photo frame that displays random photo from my Google Photo library enriched with an Haiku generated by a LLaVa instance using the same photo as reference.

Here’s a video of the digital photo frame in action:

The frame is an old Kobo Glo which I repurposed using this great guide I found on MobileRead forum and on which I’m running a super simple script that fetches and displays whatever returned from an arbitrary endpoint every minute or so.

while true; do echo "1" > /sys...

Wikimedia has a nice API which provides the facts of the day in JSON format, something like:

{"text": "Named after Bikini Atoll, the site of the nuclear weapons test Operation Crossroads in the Marshall Islands, the modern bikini was introduced at a fashion show in Paris","year": 1946}, {...

We can take one of these events and ask ChatGPT APIs to generate the description of a black and white photo representing it, we can ask for composition details, camera and a couple of tags associated ...

Lately I’ve been spent some time using ControlNet to process some of my sketches, the results are quite interesting and show the potential of the technology even when the starting sketch is very low quality and noisy, due to being on squared paper.

Quality is subjective, true, but can we show that no matter the audience a fantasy book is always better than an horror book? If this would be the case it’d mean that even readers who prefer horror books should give higher ratings to an horror-fantasy book than an horror book of other sub-genres.

As you might expect my quick investigation shows this is not the case, if we pick the first and second most voted genres for each book on Goodreads we can show that the rating distribution on the fi...

Do fantasy books on average get better ratings than sci-fi books? The answer seems to be yes, according to this little data science project I’ve build from Goodreads data.

In this project I’ve grouped all titles in the data set by category, and then calculated the mean and standard deviation for each one of them. Then I’ve followed this good tutorial to compute the confidence interval for each category, and finally I’ve plotted mean and confidence interval (with 99% confidence) in a chart, ...

This experiment combines hsl.js and tensorflow.js to perform real time object detection from a browser. When the mouse hovers the canvas the entire stream is shown, with the detected object framed in a black box, otherwise only the parts of the stream corresponding to detected objects are displayed.

You can read more about this project on github

The post In Browser Real Time Object Detection From an HTTP Live Stream first appeared on Flying memes.