An AI allowed me to recreate the “missing pixels” to build a picture several times larger than the original; ideal for an ultra-large print.

The full description is on YLovePhoto: “Upscaling a photo with free AI“.

So, here is how to proceed (thanks to the precious information from Xanny.eth).

WSL and Linux environment

Install and setup WSL, by opening a PowerShell and typing:

wsl --install

It will take a few minutes to set up. But it’s straightforward and needs no input. Just need to reboot once at the end.

Installing Ubuntu 22 LTS on the Windows PC. It is a free application from the Microsoft Store which should install right away.

When this is done, launch Ubuntu from the start menu. It will open a terminal window and request a login and password. You should enter them (and not forget them).

LLaMA dependencies

If it is not open, open a Ubuntu terminal window.

sudo apt-get update
sudo apt install make cmake build-essential python3 python3-pip git unzip

Then,

python3 -m pip install torch numpy sentencepiece

You now have a full set of background dependencies in place.

Building LLaMA itself

It is quite simple: Type the following:

git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp
make

This should be it.

Training data parameters

The real difficulty is getting the parameters (the training data). The difficulty comes from two aspects:

1. The larger the training data you want to use, the more memory you will need to run it. The alpaca-native-weights (apparently the more powerful ones easily available – about the same quality as ChatGPT 3) require more than 16GB of DRAM (I observed a 32 GB DRAM use when running them with a bunch of other things on my computer, like a couple of browsers, a mailer program, etc.)
2. The alpaca-native-weights are about 7 billion parameters (a 4+GB file to download). But they keep moving because they appear to be subject to repeated DMCA notices (the exact license of this file seems… complicated; Quite probably open source, but this is being challenged by Meta and others). So, the best you can do is to go to Pastebin to get the BitTorrent magnet and use it to download the file.

Then, the ggml-alpaca-7b-q4.bin file needs to be delivered to the llama.cpp directory.

Running LLaMA

Let the drums roll: You only have to run the command line in Ubuntu:

./main --color -i -ins -n 512 -p "You are a helpful AI who will assist, provide information, answer questions, and have conversations." -m ggml-alpaca-7b-native-q4.bin

Here is John Smith your personal AI chat assistant.

A few more recommendations

I noticed a few things that you may want to play with after the first run.

The -p option (followed by a text string) may be critical because it is setting up the background environment of your chat AI. This is an initializing prompt, not visible to the user, but deeply influencing the rest. For example, it is similar to what Microsoft or OpenAI apply beforehand in ChatGPT or Bing, in order to “give it a personality” or “to censor it”. You can play with this to freely censor your AI, or give it added freedom.

The -n 512 option has an influence on the depth of the prediction LLaMA will use. It may make it better at writing (or not) at the possible expense of CPU power.

The -t 32 (the default value) option allows defining the number of threads used by LLaMA computations. I recommend setting it to the number of threads/cores of your CPU in order to avoid spending useless efforts.

wouldn’t it be great if we had a simple way to recognize them? Yes, of course. The simpler solution? Put a different name on each drive, but this is very small step. We know that graphical interfaces work better. Why not replace all these ugly grey icons with something ore meaningful (and more pleasant)?

A simple process

The only requirements are:

• Access to Google image search (but you can use another one, of course)
• Use of a simple image editor able to save to “Windows .ICO” file format – I use IrfanView which is very feature-rich, but simple to use and perfectly free (as in free beer).

Step-by-step:

1. Define an image which is representing the drive well,
2. Collect this picture,
3. Set it to the right format,
4. Set it up on the USB drive.

Step by step

Define the image for your icon

I recommend several possible and complementary approaches. Keep in mind that the objective is to easily recognize the USB thumb drives or the disk drives connected to the computer USB port(s). Nota bene: This works with internal hard drives too.

Solution #1: If the USB thumb drive is very recognizable, you can find on the web some good photo representing it nice and well. You can even collect a press release photo form the manufacturer if you have the manufacturer name and the product model. For example, I can look for photos of external hard drives from Western Digital, MyPassport model, red color. Google result is offering a wide choice of good pictures:

AEven if you don’t have the exact name of the thumb drive (not all of Chinese-sourced drives are named properly), it is quite easy to look for a good description like “orange USB drive”. Less precise, somewhat more error-prone, but it works astonishingly well.

Solution #2: With the same Google Image Search, look for a nice photo that would represent the disk contents (or its name). Try looking for “cinema” or “movies” or “old movies” or “1980 movies” for a hard drive you expect will hold your movie collection You will be surprised by the available choice.

Choice reduction

Faced with so many options, I recommend limiting yourself with some specific search filters. (on the Google page, press the “Tools” button and choose the following filters):

1. Only take large size images (you’ll see it is important)
2. Select public domain pictures that you can modify: “Labeled for re-use with modification” (Why steal a photo when you can use one its author wants to share?).
3. Most critical: Chose a very simple, pretty but easily recognizable picture (an icon may be very small).
4. Last recommendation: It is often better to start from an image with a white or very lightly colored background (it “works” better in most cases, as you will see).

Collect the picture

Download the image you found on Google.

Create the icon

After all, an icon is merely a square image brought down to a small size and in special ICO format. the walkthrough is rather simple, you’ll see.

Load the image in IrfanView.

Modify the size of the picture to move to a square ratio (change canvas size). You can do it by either enlarging or reducing the canvas size (choosing the right background color).

The result should be a nice square picture but you may feel that it is rather large for an icon.

As a matter of fact, size does not matter much here. The larger, the better to keep the icon looking good when changing display sizes. Now comes the time to make sure that the size is exactly a multiple of 2 (eg. 2048×2048 pixels). It is not useful to go further than 4096×4096 pixels, but do not go below 1024×1024 pixels. IrfanView will do this easily, from Image>Resize/resample... menu.

Then, you need to save the picture as .ICO file (be sure to use “Save as type” with the “ICO – Windows icon” option).

Setup on the USB drive

This icon (this ICO file) must be saved at the root of the USB thumb drive (in the  D:\ directory), preferably under the autorun.ico name (but we can accommodate any other name like Lynx.ico).

Last, copy or create a text file named autorun.inf with the following contents (also in D:\ directory), :

[autorun]
ICON="autorun.ico"


This small file will allow Windows finding the icon named autorun.ico (or any other name that you previously chose)).

Result

Now, your icons will be perfectly distinguishable. Like here with drawings, pictures, photographs, logos, etc.

Bonus: this is applicable to your company logo. If you give USB thumb drives away, better than putting any photo, install your company logo for a fully personalized thumb drive!

For years, I used Google authenticator applet on iOS. I knew that it would be a pain if my phone stopped working (or if I replaced it). The only reason I was OK with this, was that I had setup Google authenticator on two phones (I happen to have one for business and one for personal use – don’t ask why…)

Actually, replacing one of these smartphones pushed me into testing Authy as a replacement for Google. And I should have done this earlier. There is no way I will switch back to Google after this.

Authy provides:

• A cute user interface
• Shared accounts between several phones (and a fast way to switch from one old phone to a new one)
• All the needed 2FA bells and whistles (compatible with the TOTP standard)

In thirty/forty minutes, I recreated all the factors for my accounts (e.g. Google, Outlook, Evernote, Dropbox, or even Ubisoft). And this was needed because I could not transfer this knowledge from the old phone to the new (even though I had both phone working in front of me). But in three/four minutes I could extend it to my other phone. So, any time I need to replace a phone, I can extend/transfer in a matter of minutes, independently of the number of web sites are included in Authy.

I feel much safer. What would have been happening if I had lost/broken my phone with the codes from Google Authenticator? (graceful degradation, no risk of losing access to many web sites at a time in case of a single failure).

I feel much more positive about extending my use of 2FA/two-factor-authentication to  much more web sites because of the added ease of use (longer-term commitment).

Furthermore, Authy is available for several Operating Systems:

• iOS
• Android
• Windows
• Google Chrome

Go and install it!