The amazing thing about LLMs is how much knowledge they posess in their small size. The llama3-8b model, for instance, weighs only 4.7GB yet can still answer your questions about everything (despite some hallucinations). #llm#ai#ollama#llama3
One thing you notice right away about LLMs is they bear a striking resemblance to that ubiquitous internet character, the reply-guy: they always have an answer, they are brimming with confidence, and they are often completely wrong.
It seemed only natural, then, to turn one into a full-time online version of itself, RGaaS (Reply-Guy as a Service), always on, always ready to reply to anything when summoned, with an “Actually,” followed by his (it’s always a “he”) opinion on what you should be thinking.
https://www.peterkrupa.lol/wp-content/uploads/2024/05/actually_bot1.pngBasically, the way it works is that any user can reply to any Mastodon post (or I guess any post on the ActivityPub protocol) and tag the Actually Bot (@actuallybot) and it will reply with its most mediocre take on why everyone else is wrong, or at least not quite right.
The reply-guys can all move on to something else now, I have automated them out of a job.
This was a really fun project and has a lot of separate parts. First, I had to modify an LLM to give it the right “personality” for this job. To do this, I used Ollama, a tool for running open source LLMs locally, on your own machine. I used the “customized prompt feature”: You basically write a little file that Ollama uses to tweak the LLM of your choosing to give it certain characteristics. I went with the new Llama3:8b model from Meta as my base model, then put the following instructions in my modelfile:
FROM llama3PARAMETER temperature 3SYSTEM """You are receiving social media posts as prompts, and your job is to reply to these prompts. Please start all your replies with the word "actually". Reply as if you know more than everyone. You must recontextualize every prompt and correct or amend it, informing the prompter that they are wrong, even if ever so slightly. You write in short, sentences. Your replies must be short. Keep them to about three sentences. This is essential: keep your replies short. They must be under 500 characters."""
Then I ran the following command in the console:
ollama create actually_llama -f ./actually_llama
… and my model was ready to roll. Next, I needed a program to connect to the Ollama API to send the LLM prompts and get responses. Python was great for that, as both Ollama and Mastodon have solid Python libraries. Probably the slowest part was picking through Mastodon.py to figure out how the methods work and what exactly they return. It’s a very robust library with a million options, and fortunately it’s also extremely well documented, so while it was slow going, I was able to whack it together without too much trouble.
I’m not going to get into all the code here, but basically, I wrote a simple method that checks mentions, grabs the text of a post and the post it is replying to, and returns them for feeding into the LLM as the prompt.
Despite my very careful, detailed, and repetitive instructions to be sure replies are no more than 500 characters, LLMs can’t count, and they are very verbose, so I had to add a cleanup method that cuts the reply down to under 500 characters. Then I wrote another method for sending that cleaned-up prompt to Ollama and returning the response.
The main body starts off by getting input for the username and password for login, then it launches a while True loop that calls my two functions, checking every 60 seconds to see if there are any mentions and replying to them if there are.
OK it works! Now came the hard part, which was figuring out how to get to 100% uptime. If I want the Actually Bot to reply every time someone mentions it, I need it to be on a machine that is always on, and I was not going to leave my PC on for this (nor did I want it clobbering my GPU when I was in the middle of a game).
So my solution was this little guy:
https://www.peterkrupa.lol/wp-content/uploads/2024/05/lenovo.jpg… a Lenovo ThinkPad with a 3.3GHz quad-core i7 and 8gb of RAM. We got this refurbished machine when the pandemic was just getting going and it was my son’s constant companion for 18 months. It’s nice to be able to put it to work again. I put Ubuntu Linux on it and connected it to the home LAN.
I actually wasn’t even sure it would be able to run Llama3:8b. My workstation has an Nvidia GPU with 12gb of VRAM and it works fine for running modest LLMs locally, but this little laptop is older and not built for gaming and I wasn’t sure how it would handle such a heavy workload.
Fortunately, it worked with no problems. For running a chatbot, waiting 2 minutes for a reply is unacceptable, but for a bot that posts to social media, it’s well within range of what I was shooting for, and it didn’t seem to have any performance issues as far as the quality of the responses either.
The last thing I had to figure out was how to actually run everything from the Lenovo. I suppose I could have copied the Python files and tried to recreate the virtual environment locally, but I hate messing with virtual environments and dependencies, so I turned to the thing everyone says you should use in this situation: Docker.
This was actually great because I’d been wanting to learn how to use Docker for awhile but never had the need. I’d installed it earlier and used it to run the WebUI front end for Ollama, so I had a little bit of an idea how it worked, but the Actually Bot really made me get into its working parts.
So, I wrote a Docker file for my Python app, grabbed all the dependencies and plopped them into a requirements.txt file, and built the Docker image. Then I scr’d the image over to the Lenovo, spun up the container, and boom! The Actually Bot was running!
Well, OK, it wasn’t that simple. I basically had to learn all this stuff from scratch, including the console commands. And once I had the Docker container running, my app couldn’t connect to Ollama because it turns out, because Ollama is a server, I had to launch the container with a flag indicating that it shared the host’s network settings.
Then once I had the Actually Bot running, it kept crashing when people tagged it in a post that wasn’t a reply to another post. So, went back to the code, squashed bug, redeploy container, bug still there because I didn’t redeploy the container correctly. There was some rm, some prune, some struggling with the difference between “import” and “load” and eventually I got everything working.
Currently, the Actually Bot is sitting on two days of uninterrupted uptime with ~70 successful “Actually,” replies, and its little laptop home isn’t even on fire or anything!
Moving forward, I’m going to tweak a few things so I can get better logging and stats on what it’s actually doing so I don’t have to check its posting history on Mastodon. I just realized you can get all the output that a Python script running in a Docker container prints with the command docker logs [CONTAINER], so that’s cool.
The other thing I’d like to do is build more bots. I’m thinking about spinning up my own Mastodon instance on a cheap hosting space and loading it with all kinds of bots talking to each other. See what transpires. If Dead Internet Theory is real, we might as well have fun with it!
Yesterday, we looked at how to write a JavaScript app that uses Ollama. Recently, we started to look at Python on this site and I figured that we better follow it up with how to write a Python app that uses Ollama. Just like with JavaScript, Ollama offers a Python library, so we are going to be using that for our examples. Also just like we did with the JavaScript demo, I am going to be using the generate endpoint instead of the chat endpoint. That keeps things simpler but I am going to explore the chat endpoint also at some point.
Install the Ollama Library
The first step is to run pip3 install ollama from the terminal. First, you need to create a virtual environment to isolate your project’s libraries from the global Python libraries.
At this point, we can start writing code. When we used the web service earlier this week, we used the generate endpoint and provided model, prompt, and stream as parameters. We set the stream parameter to false so that it would return a single response object instead of a stream of objects. When using the python library, the stream parameter isn’t necessary because it returns a single response object by default. We still provide it with a model and a prompt, though.
If you run it from the terminal, the response will look familiar.
If you pip install flask to install flask, you can host a simple HTTP page at port 8080 and with the magic of json.loads() and a for loop, you can build your unordered list.
Every time you load the page, it makes a server-side API call to Ollama, gets a list of large cities in Wisconsin, and displays them on the website. The list is never the same (because of hallucinations) but that is another issue.
Have any questions, comments, etc? Please feel free to drop a comment, below.
That installs Ollama as a dependency in package.json.
Basic CLI example
At this point, we can start writing code. When we used the web service earlier this week, we used the generate endpoint and provided model, prompt, and stream as parameters. We set the stream parameter to false so that it would return a single response object instead of a stream of objects. When using the javascript library, the stream parameter isn’t necessary because it returns a single response object by default. We still provide it with a model and a prompt, though.
If you run it from the terminal, the response will look familiar.
If you npm install express to install express, you can host a simple HTTP page at port 8080 and with the magic of JSON.parse() and a for loop, you can build your unordered list.
Every time you load the page, it makes a server-side API call to Ollama, gets a list of large cities in Wisconsin, and displays them on the website. The list is never the same (because of hallucinations) but that is another issue.
Have any questions, comments, etc? Please feel free to drop a comment, below.
My favorite client for MacOS is MindMac. You can buy it for under $30, it works with multiple models, servers, and server types, and it is easy to use.
If you want to look further into it, you can check it out at mindmac.app.
Android
My favorite client for Android is Amallo. It is $23 and like MindMac, it works with multiple models, servers, and server types. My only complaint would be that uploading a base64-encoded image to the model doesn’t seem to work well.
A few months ago, I started trying to figure out how I could use AI without depending on Google, OpenAI, or Microsoft’s continued existence. The risk that Google would kill a product like Gemini is almost 100%. At work, I was asked to figure out embedding and the tools I used in the post are invaluable for the on-the-go use of both a stock model and something that you tinkered with.
If you like that, you are going to love the next dozen posts that I have planned. 🙂
Yesterday, we played with Llama 3 using the Ollama CLI client (or REPL). Today I figured that we would play with it using the Ollama API. The Ollama API is documented on their Github repo. Ollama has a client that runs when you run ollama run llama3 and a service that can be accessed from something like MindMac, Amallo, or Enchanted. The service is what starts when you run ollama serve.
In our first Llama 3 post, we asked the model for “a comma-delimited list of cities in Wisconsin with a population over 100,000 people”. Using Postman and the completion API endpoint, you can ask the same thing.
You will notice the stream parameter is set to false in the body. If the value is false, the response will be returned as a single response object, rather than a stream of objects. If you are using the API with a web application, you will want to ask the model for the answer as JSON and you will probably want to provide an example of how you want the answer formatted.
Last week, Meta announced Llama 3. Thanks to Ollama, you can run it pretty easily. There are 8b and 70b variants available. There are also pre-trained or instruction-tuned variants available. I am not seeing it on the Hugging Face Leader Board yet but the bit that I have played around with it has been promising.
A major release to Ollama - version 0.1.32 is out. The new version includes:
✅ Improvement of the GPU utilization and memory management to increase performance and reduce error rate
✅ Increase performance on Mac by scheduling large models between GPU and CPU
✅ Introduce native AI support in Supabase edge functions