Using Local AI On The Command Line To Rename Images (And More)

We all have a folder full of images whose filenames resemble line noise. How about renaming those images with the help of a local LLM (large language model) executable on the command line? All that and more is showcased on [Justine Tunney]’s bash one-liners for LLMs, a showcase aimed at giving folks ideas and guidance on using a local (and private) LLM to do actual, useful work.

This is built out from the recent llamafile project, which turns LLMs into single-file executables. This not only makes them more portable and easier to distribute, but the executables are perfectly capable of being called from the command line and sending to standard output like any other UNIX tool. It’s simpler to version control the embedded LLM weights (and therefore their behavior) when it’s all part of the same file as well.

One such tool (the multi-modal LLaVA) is capable of interpreting image content. As an example, we can point it to a local image of the Jolly Wrencher logo using the following command:

llava-v1.5-7b-q4-main.llamafile --image logo.jpg --temp 0 -e -p '### User: The image has...\n### Assistant:'

Which produces the following response:

The image has a black background with a white skull and crossbones symbol.

With a different prompt (“What do you see?” instead of “The image has…”) the LLM even picks out the wrenches, but one can already see that the right pieces exist to do some useful work.

Check out [Justine]’s script, which cleverly evaluates image filenames. If an image’s given filename already looks like readable English (also a job for a local LLM) the image is left alone. Otherwise, the picture is fed to an LLM whose output guides the generation of a new short and descriptive English filename in lowercase, with underscores for spaces.

What about the fact that LLM output isn’t entirely predictable? That’s easy to deal with. [Justine] suggests always calling these tools with the --temp 0 parameter. Setting the temperature to zero makes the model deterministic, ensuring that a same input always yields the same output.

There’s more neat examples on the Bash One-Liners for LLMs that demonstrate different ways to use a local LLM that lives in a single-file executable, so be sure to give it a look and see if you get any new ideas. After all, we have previously shown how automating tasks is almost always worth the time invested.

AI Assistant Translates Your Every Request For The Command Line

If you don’t live on the command line, it can be easy to forget the exact syntax of commands. It often leaves you running to the “/?” or “–help” switches, or else a quick Google search to find the proper incantations. Shell-AI is a machine-learning assistant that could change all that by helping you find the proper command for the job, right on the command line!

Shell-AI accepts natural-language inputs — simply type in “shai” followed by what you’re trying to do. It will then take in your request, run it through an OpenAI language model like GPT-3.5-Turbo, and then present you with three (or more) potential commands. You can then select which command to use and get on with your day.

As demonstrated, it’s more than capable of following commands like “download a random image” or “show only image files ls.” And, hilariously, it responds to the request “do something crazy” with just one suggestion: “rm -rf”. That seems rather fitting.

We wouldn’t blindly follow any commands coming out of a large language model, of course. But, if you know what you’re doing, it could prove a useful little tool to ease your regular duties on the command line.

Linux Fu: Making Progress

The computer world looks different from behind a TeleType or other hardcopy terminal. Things that tend to annoy people about Unix or Linux these days were perfectly great when you were printing everything the computer said to you. Consider the brevity of most basic commands. When you copy a file, for example, it doesn’t really tell you much other than it returns you to the prompt when it is done. If you are on a modern computer working with normal-sized files locally, not a big deal. But if you are over a slow network or with huge files, it would be nice to have a progress bar. Sure, you could write your own version of copy, but wouldn’t it be nice to have some more generic options?

One Way

The pv program can do some of the things you want. It monitors data through a pipe or, at least through its standard output. Think of it as cat with a meter. Suppose you want to write a diskimage to /dev/sdz:

cat diskz.img >/dev/sdz

But you could also do:

pv diskz.img >/dev/sdz

By default, pv will show a progress bar, an elapsed time, an estimated end time, a rate, and a total number of bytes. You can turn any of that off or add things using command line options. You can also specify things like the size of the terminal if it should count lines instead of bytes, and, in the case where the program doesn’t know what it is reading, the expected size of the transfer.

Continue reading “Linux Fu: Making Progress”

CoreFreq Gives Peek At CPU Performance Info On Linux

The CPU is the part of the computer that makes everything else tick. While GPUs have increasingly become a key part of overall system performance, we still find ourselves wanting to know how our CPU is doing. CoreFreq is a Linux tool that aims to tell you everything you want to know about your modern 64-bit CPU.

The tool relies on a kernel module, and is coded primarily in C, with some assembly code used to measure performance as accurately as possible. It’s capable of reporting everything from core frequencies to details on hyper-threading and turbo boost operation. Other performance reports include information on instructions per cycle or instructions per second, and of course, all the thermal monitoring data you could ask for. It all runs in the terminal, which helps keep overheads low.

The hardcore among us can build it from source, available on GitHub, though it’s reportedly available in package form, and as a live CD, too. We could imagine data captured from CoreFreq could be used for some fun performance visualizations, too. If you’ve been whipping up your own nifty command-line tools, be sure to drop us a line!

Cat9 And LASH Want To Change Your Linux Command Line

It is no secret that to be a true Linux power user you have to deal with the command line. Many people actually prefer to use the command line. However, the shell — the program that provides that command line — is mired in a back history which means it has to work with existing things no matter how modern it tries to be. However, a new set of projects wants to replace most of your user interface stack starting with the shell. At the top of that stack is Cat9 which is technically a shell, but not in the way you probably imagine a shell.

A traditional shell lets you run programs one at a time, feed them input, and observe their output. Sure, you can stash the output away for later use. You can run programs in the background or in parallel, but that requires special attention. In Cat9, everything is asynchronous and results stay around until you deliberately drop them. It is trivial to grab data from a previous command or, for example, to switch to a directory that was in use by an earlier task.

Continue reading “Cat9 And LASH Want To Change Your Linux Command Line”

Building A Serial Bus To Save An Old Hard Drive

Universal Serial Bus has been the de facto standard for sending information to and from computer peripherals for almost two decades, but despite the word “universal” in the name this wasn’t always the case. Plenty of competing standards, including USB, existed in the computing world in the decades before it came to dominance, and if you’re trying to recover data from a computer without USB you might have to get creative with how it’s done.

[Ben] recently came across a 80486 with this problem, so he had to get creative to recover the contents of the drive. He calls it the “lunchbox” computer due to its form factor, and while it doesn’t have USB it does have a tried-and-trusted serial port to communicate with other computers. [Ben] wrote up a piece of software for both the receiving computer and the sending computer in order to copy the drive sectors one by one across a serial link to a standalone computer running Windows XP, and was able to recover the contents of the drive that way instead.

All of the code [Ben] wrote is available on his GitHub page for anyone looking to boot up a 30-year-old computer again. While it might sound uncommon, computers of this vintage are still around running things like CNC machines or old mainframes.

Remote Screen Viewer Is Text-Only

Have you been slowly falling down a rabbit hole of Stallman-like paranoia of computers ever since installing Ubuntu for the first time in 2007? Do you now abhor anything with a GUI, including browsers? Do you check your mail with the command line even though you’re behind seven proxies? But, do you still want to play Minecraft? If so, this command-line-only screen viewer might just be the tool to use a GUI without technically using one.

This remote screen viewer is built in Python by [louis-e] and, once installed, allows the client to view the screen of the server even if the client is a text-only console. [louis-e] demonstrates this from within a Windows command prompt. The script polls the server screen and then displays it in the console using the various colors and textures available. As a result, the resolution and refresh rate are both quite low, but it is still functional enough to play Minecraft and do other GUI-based tasks as long as there’s no fine text to read anywhere.

The video below only shows a demonstration of the remote screen viewer, and we can imagine plenty of uses beyond this proof-of concept game demonstration. Installing a desktop environment and window manager is not something strictly necessary for all computers, so this is a functional workaround if you don’t want to waste time and resources installing either of those components. If you’re looking for remote desktop software for a more specific machine, though, take a look at this software which enables remote desktop on antique Macs.

Continue reading “Remote Screen Viewer Is Text-Only”