843 points | by armanified 1 day ago
79 comments
- thomasfl 12 hours agoIs there some documentation for this? The code is probably the simplest (Not So) Large Language Model implementation possible, but it is not straight forward to understand for developers not familiar with multi-head attention, ReLU FFN, LayerNorm and learned positional embeddings.
This projects shares similarities with Minix. Minix is still used at universities as an educational tool for teaching operating system design. Minix is the operating system that taught Linus Torvalds how to design (monolithic) operating systems. Similarly having students adding capabilities to GuppyLM is a good way to learn LLM design.
- achenatx 12 hours agogive the code to an LLM and have a discussion about it.
- dominotw 10 hours agodoes this work? there is no more need for writing high level docs?
- arcanemachiner 10 hours ago> does this work?
Absolutely. If you loaded this into an agentic coding harness with a decent model, I can practically guarantee it would be able to help you figure out what's going on.
> there is no more need for writing high level docs?
Absolutely not. That would be like exploring a cave without a flashlight, knowing that you could just feel your way around in the dark instead.
Code is not always self-documenting, and can often tell you how it was written, but not why.
- stronglikedan 8 hours ago> If you loaded this into an agentic coding harness with a decent model, I can practically guarantee it would be able to help you figure out what's going on.
My non-coder but technically savvy boss has been doing this lately to great success. It's nice because I spend less time on it since the model has taken my place for the most part.
- libria 6 hours ago> since the model has taken my place for the most part
Hah, you realize the same thing is going on in your boss's head right? The pie chart of Things-I-Need-stronglikedan-For just shrank tiny bit...
- dominotw 4 hours agomy last employer was using ai to rank developers on most impactful code their prs are shipping.
- sigmoid10 10 hours agoThere are so many blogs and tutorials about this stuff in particular, I wouldn't worry about it being outside the training data distribution for modern LLMs. If you have a scarce topic in some obscure language I'd be more careful when learning from LLMs.
- bigmadshoe 10 hours agoLLMs can tell you what the code does but not why the developer chose to do it that way.
Also, large codebases are harder to understand. But projects like these are simple to discuss with an LLM.
- stronglikedan 8 hours ago> LLMs can tell you what the code does but not why the developer chose to do it that way.
Do LLMs not take comments into consideration? (Serious question - I'm just getting into this stuff)
- dr_hooo 7 hours agoThey do (it's just text), if they are there...
- fg137 15 hours agoHow does this compare to Andrej Karpathy's microgpt (https://karpathy.github.io/2026/02/12/microgpt/) or minGPT (https://github.com/karpathy/minGPT)?
- armanified 14 hours agoI haven't compared it with anything yet. Thanks for the suggestion; I'll look into these.
- BrokenCogs 12 hours agoWho cares how it compares, it's not a product it's a cool project
- tantalor 12 hours agoEven cool projects can learn from others. Maybe they missed something that could benefit the project, or made some interesting technical choice that gives a different result.
For the readers/learners, it's useful to understand the differences so we know what details matter, and which are just stylistic choices.
This isn't art; it's science & engineering.
- BrokenCogs 12 hours agoBut it isn't the OP's responsibility to compare their project to all other projects. The GP could themselves perform the comparison and post their thoughts instead of asking an open ended question.
- philipallstar 11 hours ago> it isn't the OP's responsibility to compare their project to all other projects
No one, including the GP, said it was.
- fg137 11 hours agoIt isn't, but such information will be immensely helpful to anyone who wants to learn from such projects. Some tutorials are objectively better than others, and learners can benefit from such information.
- tantalor 12 hours ago100% agree, I didn't mean to imply that OP is responsible for that, or that the (lack of) comparison detracts in any way from the work.
- stronglikedan 8 hours ago> Who cares how it compares
Well, the person who asked the question, for one. I'm sure they're not the only one. Best not to assume why people are asking though, so you can save time by not writing irrelevant comments.
- layer8 10 hours agoMicrogpt isn’t a product either. Are you saying that differences between cool projects aren’t worth thinking and conversing about?
- totetsu 17 hours agohttps://bbycroft.net/llm has 3d Visualization of tiny example LLM layers that do a very good job at showing what is going on (https://news.ycombinator.com/item?id=38505211)
- armanified 14 hours agoPretty neat! I'll definitely take a deeper look into this.
- maverickxone 15 hours agohave little to do with this, but i have to say your project are indeed pretty cool! Consider adding some more UI?
- skramzy 12 hours agoNeat!
- ordinarily 23 hours agoIt's genuinely a great introduction to LLMs. I built my own awhile ago based off Milton's Paradise Lost: https://www.wvrk.org/works/milton
- mudkipdev 20 hours agoThis is probably a consequence of the training data being fully lowercase:
You> hello Guppy> hi. did you bring micro pellets.
You> HELLO Guppy> i don't know what it means but it's mine.
- functional_dev 19 hours agoGreat find! It appears uppercase tokens are completely unknonw to the tokenizer.
But the character still comes through in response :)
- algoth1 14 hours agoThis really makes me think if it would be feasible to make an llm trained exclusively on toki pona (https://en.wikipedia.org/wiki/Toki_Pona)
- MarkusQ 11 hours agoThere isn't enough training data though, is there? The "secret sauce" of LLMs is the vast amount of training data available + the compute to process it all.
- algoth1 9 hours agoI think you could probably feed a copy of a toki pona grammar book to a big model, and have it produce ‘infinite’ training data
- MarkusQ 6 hours agoThis is essentially a distillation on the bigger model; you'd wind up surfacing a lot of artifacts from the host model, amplifying them in the same way repeated photocopying introduces errors.
https://dailyai.com/2025/05/create-a-replica-of-this-image-d...
- eden-u4 7 hours agoThere are not enough samples in that book to generate new "infinite" data.
- hackerman70000 17 hours agoFinally an LLM that's honest about its world model. "The meaning of life is food" is arguably less wrong than what you get from models 10,000x larger
- amelius 14 hours agoIt's arguably even better than the most famous answer to that question.
- zkmon 14 hours agoMeaning/goal of life is to reproduce. Food (and everything else) is only a means to it. Reproduction is the only root goal given by nature to any life form. All resources and qualities are provided are only to help mating.
- tantalor 12 hours agoReproduction is the goal of genes.
Food (not dying) is the goal of organisms.
- philote 11 hours agoI'd argue genes nor life has a "goal". They are what they are because they've been successful at continuing their existence. Would you say a rock's goal is not to get broken?
- tantalor 11 hours agoOnly because genes/organisms can make choices (changes to its programming, or decisions) to optimize their path towards their goal.
A rock is maybe not a good counterexample, but a crystal is because it can grow over time. So in some sense, it tries not to break. However a crystal cannot make any choices; it's behavior is locked into the chemistry it starts with.
- amelius 14 hours agoThen why are reproductive rates so low in western countries?
https://en.wikipedia.org/wiki/List_of_countries_by_total_fer...
- darepublic 13 hours agoThe western lifestyle is an evolutionary dead end?
- vixen99 13 hours agoIt seems that some in the West want it to be and are working hard to make it so.
- michaelhoney 4 hours agonot just western countries
- hca 11 hours agoNo, evolution has encoded lust. It has not yet allowed for condoms. But it's a process.
- neurworlds 12 hours agoCool project. I'm working on something where multiple LLM agents share a world and interact with each other autonomously. One thing that surprised me is how much the "world" matters — same model, same prompt, but put it in a system with resource constraints, other agents, and persistent memory, the behavior changes dramatically. Made me realize we spend too much time optimizing the model and not enough thinking about the environment it operates in.
- zwaps 20 hours agoI like the idea, just that the examples are reproduced from the training data set.
How does it handle unknown queries?
- armanified 14 hours agoIt mostly doesn't, at 9M it has very limited capacity. The whole idea of this project is to demonstrate how Language Models work.
- bblb 17 hours agoCould it be possible to train LLM only through the chat messages without any other data or input?
If Guppy doesn't know regular expressions yet, could I teach it to it just by conversation? It's a fish so it wouldn't probably understand much about my blabbing, but would be interesting to give it a try.
Or is there some hard architectural limit in the current LLM's, that the training needs to be done offline and with fairly large training set.
- BiraIgnacio 8 hours agoNice work and thanks for sharing it!
Now, I ask, have LLMs ben demystified to you? :D
I am still impressed how much (for the most part) trivial statistics and a lot of compute can do.
- rpdaiml 11 hours agoThis is a nice idea. A tiny implementation can be way more useful for learning than yet another wrapper around a big model, especially if it keeps the training loop and inference path small enough to read end to end.
- ergocoder 5 hours agoIt's just so amazing that 5 years ago it would be extremely to build a conversational bot like this.
But right now people make it a hobby, and that thing can run on a laptop.
This is just so wild.
- cbdevidal 22 hours ago> you're my favorite big shape. my mouth are happy when you're here.
Laughed loudly :-D
- vunderba 22 hours agoThis is a direct output from the synthetic training data though - wonder if there is a bit of overfitting going on or it’s just a natural limitation of a much smaller model.
- Leomuck 11 hours agoWow that is such a cool idea! And honestly very much needed. LLMs seem to be this blackbox nobody understands. So I love every effort to make that whole thing less mysterious. I will definitely have a look at dabbling with this, may it not be a goldfish LLM :)
- CaseFlatline 11 hours agoI am trying to find how the synthetic data was created (looking through the repo) and didn't find it. Maybe I am missing it - Would love to see the prompts and process on that aspect of the training data generation!
- vunderba 10 hours agoIt's here:
https://github.com/arman-bd/guppylm/blob/main/guppylm/genera...
Uses a sort of mad-libs templatized style to generate all the permutations.
- bharat1010 8 hours agoThis is such a smart way to demystify LLMs. I really like that GuppyLM makes the whole pipeline feel approachable..great work
- kaipereira 20 hours agoThis is so cool! I'd love to see a write-up on how made it, and what you referenced because designing neural networks always feel like a maze ;)
- jzer0cool 11 hours agoDoes this work by just training once with next token prediction? Want to understand better how it creates fluent sentences if anyone can provide insights.
- brcmthrowaway 20 hours agoWhy are there so many dead comments from new accounts?
- 59nadir 17 hours agoBecause despite what HN users seem to think, HN is a LLM-infested hellscape to the same degree as Reddit, if not more.
- wiseowise 15 hours agoYou’re absolutely right! HN isn’t just LLM-infested hellscape, it’s a completely new paradigm of machine assisted chocolate-infused information generation.
- toyg 15 hours agoJust let me know which type of information goo you'd like me to generate, and I'll tailor the perfect one for you.
- siva7 14 hours agoBut what should we do? The parent company isn't transparent about communicating the seriousness of this problem
- loveparade 18 hours agoIt really seems it's mostly AI comments on this. Maybe this topic is attractive to all the bots.
- armanified 13 hours agoThis title might have triggered something in those bots; most of them have sneaky AI SaaS links in their bio.
Honestly, I never expected this post to become so popular. It was just the outcome of a weekend practice session.
- AlecSchueler 20 hours agoThey all seem to be slop comments.
- ankitsanghi 20 hours agoLove it! I think it's important to understand how the tools we use (and will only increasingly use) work under the hood.
- Duplicake 15 hours agoI love this! Seems like it can't understand uppercase letters though
- armanified 14 hours agoUppercase letters were intentionally ignored.
- NyxVox 22 hours agoHm, I can actually try the training on my GPU. One of the things I want to try next. Maybe a bit more complex than a fish :)
- EmilioOldenziel 9 hours agoBuilding it yourself is always the best test if you really understand how it works.
- drincanngao 14 hours agoI was going to suggest implementing RoPE to fix the context limit, but realized that would make it anatomically incorrect.
- armanified 14 hours agoI intentionally removed all optimizations to keep it vanilla.
- jbethune 9 hours agoForked. Very cool. I appreciate the simplicity and documentation.
- nobodyandproud 11 hours agoThanks. Tinkering is how I learn and this is what I’ve been looking for.
- fawabc 15 hours agohow did you generate the synthetic data?
- SilentM68 23 hours agoWould have been funny if it were called "DORY" due to memory recall issues of the fish vs LLMs similar recall issues :)
- armanified 14 hours agoOMG! Why didn't I thought fo this first :P
- kubrador 20 hours agohow's it handle longer context or does it start hallucinating after like 2 sentences? curious what the ceiling is before the 9M params
- amelius 14 hours ago> A 9M model can't conditionally follow instructions
How many parameters would you need for that?
- armanified 14 hours agoMy initial idea was to train a navigation decision model with 25M parameters for a Raspberry Pi, which, in testing, was getting about 60% of tool calls correct. IMO, it seems like around 20M parameters would be a good size for following some narrow & basic language instructions.
- amelius 13 hours agoOk. This makes me wonder about a broader question. Is there a scientific approach showing a pyramid of cognitive functions, and how many parameters are (minimally) required for each layer in this pyramid?
- gnarlouse 22 hours agoI... wow, you made an LLM that can actually tell jokes?
- murkt 18 hours agoWith 9M params it just repeats the joke from a training dataset.
- ben8bit 17 hours agoThis is really great! I've been wanting to do something similar for a while.
- rclkrtrzckr 19 hours agoI could fork it and create TrumpLM. Not a big leap, I suppose.
- search_facility 17 hours agoprobably 8M params are too much even :)
- danparsonson 15 hours agoAs long as you use the best parameters then it doesn't matter
- wiseowise 15 hours agoGrab her by the pointer.
- winter_blue 10 hours agoThis is amazing work. Thank you.
- rahen 12 hours agoI don't mean to be 'that guy', but after a quick review, this really feels like low-effort AI slop to me.
There is nothing wrong using AI tools to write code, but nothing here seems to have taken more than a generic 'write me a small LLM in PyTorch' prompt, or any specific human understanding.
The bar for what constitutes an engineering feat on HN seems to have shifted significantly.
- ananandreas 15 hours agoGreat and simple way to bridge the gap between LLMs and users coming in to the field!
- cpldcpu 18 hours agoLove it! Great idea for the dataset.
- monksy 19 hours agoIs this a reference from the Bobiverse?
- nullbyte808 1 day agoAdorable! Maybe a personality that speaks in emojis?
- armanified 14 hours agoOMG! You just gave me the next idea..
- AndrewKemendo 1 day agoI love these kinds of educational implementations.
I want to really praise the (unintentional?) nod to Nagel, by limiting capabilities to representation of a fish, the user is immediately able to understand the constraints. It can only talk like a fish cause it’s very simple
Especially compared to public models, thats a really simple correspondence to grok intuitively (small LLM > only as verbose as a fish, larger LLM > more verbose) so kudos to the author for making that simple and fun.
- dvt 23 hours ago> the user is immediately able to understand the constraints
Nagel's point was quite literally the opposite[1] of this, though. We can't understand what it must "be like to be a bat" because their mental model is so fundamentally different than ours. So using all the human language tokens in the world can't get us to truly understand what it's like to be a bat, or a guppy, or whatever. In fact, Nagel's point is arguably even stronger: there's no possible mental mapping between the experience of a bat and the experience of a human.
[1] https://www.sas.upenn.edu/~cavitch/pdf-library/Nagel_Bat.pdf
- Terr_ 19 hours agoIMO we're a step before that: We don't even have a real fish involved, we have a character that is fictionally a fish.
In LLM-discussions, obviously-fictional characters can be useful for this, like if someone builds a "Chat with Count Dracula" app. To truly believe that a typical "AI" is some entity that "wants to be helpful" is just as mistaken as believing the same architecture creates an entity that "feels the dark thirst for the blood of the living."
Or, in this case, that it really enjoys food-pellets.
- andoando 19 hours agoId highly disagree with that. Were all living in the same shared universe, and underlying every intelligence must be precisely an understanding of events happening in this space-time.
- vixen99 12 hours agoWhat does 'precisely' mean? Everyone has the same understanding of events - a precise one?
- andoando 9 hours agoNo I am saying the basis of intelligence must be shared, not that we have the same exact mental model.
I might for example say a human entered a building, a bat might on the other hand think "some big block with two sticks moved through a hole", but both are experiencing a shared physical observation, and there is some mapping between the two.
Its like when people say, if there are aliens they would find the same mathematical constants thet we do
- AndrewKemendo 23 hours agoDifferent argument
I’m not going to argue other than to say that you need to view the point from a third party perspective evaluating “fish” vs “more verbose thing,” such that the composition is the determinant of the complexity of interaction (which has a unique qualia per nagel)
Hence why it’s a “unintentional nod” not an instantiation
- Vektorceraptor 12 hours agoHaha, funny name :)
- gdzie-jest-sol 17 hours ago* How creating dataset? I download it but it is commpresed in binary format.
* How training. In cloud or in my own dev
* How creating a gguf
- freetonik 17 hours agoYou sound like Guppy. Nice touch.
- gdzie-jest-sol 17 hours ago``` uv run python -m guppylm chat
Traceback (most recent call last):
Exception: No such file or directory (os error 2) ```File "<frozen runpy>", line 198, in _run_module_as_main File "<frozen runpy>", line 88, in _run_code File "/home/user/gupik/guppylm/guppylm/__main__.py", line 48, in <module> main() File "/home/user/gupik/guppylm/guppylm/__main__.py", line 29, in main engine = GuppyInference("checkpoints/best_model.pt", "data/tokenizer.json") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/user/gupik/guppylm/guppylm/inference.py", line 17, in __init__ self.tokenizer = Tokenizer.from_file(tokenizer_path) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^- gdzie-jest-sol 16 hours agomeybe add training again (read best od fine) and train again
``` # after config device checkpoint_path = "checkpoints/best_model.pt"
ckpt = torch.load(checkpoint_path, map_location=device, weights_only=False)
model = GuppyLM(mc).to(device) if "model_state_dict" in ckpt: model.load_state_dict(ckpt["model_state_dict"]) else: model.load_state_dict(ckpt)
start_step = ckpt.get("step", 0) print(f"Encore {start_step}") ```
- hughw 13 hours agoTiny LLM is an oxymoron, just sayin.
- uxcolumbo 13 hours agoHow about: LLMs are on a spectrum and this one is on the tiny side?
- armanified 13 hours agoTrue, but most would ignore LM if it weren't LLM.
- oyebenny 20 hours agoNeat!
- Elengal 16 hours agoCool
- meidad_g 9 hours ago[dead]
- techpulselab 9 hours ago[dead]
- textai2026 12 hours ago[dead]
- zephyrwhimsy 4 hours ago[dead]
- adamsilvacons 12 hours ago[dead]
- maxothex 10 hours ago[dead]
- Morpheus_Matrix 23 hours ago[flagged]
- solsafe_dev 11 hours ago[dead]
- agdexai 12 hours ago[dead]
- ethanmacavoy 22 hours ago[flagged]
- zephyrwhimsy 8 hours ago[dead]
- Alexzoofficial 19 hours ago[flagged]
- agenexus 22 hours ago[flagged]
- peifeng07 18 hours ago[dead]
- zephyrwhimsy 18 hours ago[dead]
- zephyrwhimsy 13 hours ago[dead]
- novachen 14 hours ago[dead]
- weiyong1024 23 hours ago[flagged]
- techpulselab 17 hours ago[dead]
- aesopturtle 21 hours ago[flagged]
- Morpheus_Matrix 17 hours ago[dead]
- solsafe_dev 11 hours ago[dead]
- george_belsky 18 hours ago[dead]
- aditya7303011 21 hours agoDid something similar last year https://github.com/aditya699/EduMOE
- zhichuanxun 14 hours ago[dead]
- aditya7303011 21 hours ago[dead]
- LeonTing1010 22 hours ago[flagged]
- secabeen 22 hours agoTraining data is here:
https://huggingface.co/datasets/arman-bd/guppylm-60k-generic
- jiusanzhou 19 hours ago[flagged]
- areys 13 hours ago[flagged]
- moonu 13 hours agoThis comment seems ai-written
- dinkumthinkum 21 hours agoI think this is a nice project because it is end to end and serves its goal well. Good job! It's a good example how someone might do something similar for a specific purpose. There are other visualizers that explain different aspects of LLMs but this is a good applied example.
- martmulx 22 hours agoHow much training data did you end up needing for the fish personality to feel coherent? Curious what the minimum viable dataset looks like for something like this.
- Propelloni 17 hours agoGreat work! I still think that [1] does a better job of helping us understand how GPT and LLM work, but yours is funnier.
Then, some criticism. I probably don't get it, but I think the HN headline does your project a disservice. Your project does not demystify anything (see below) and it diverges from your project's claim, too. Furthermore, I think you claim too much on your github. "This project exists to show that training your own language model is not magic." and then just posts a few command line statements to execute. Yeah, running a mail server is not magic, just apt-get install exim4. So, code. Looking at train_guppylm.ipynb and, oh, it's PyTorch again. I'm better off reading [2] if I'm looking into that (I know, it is a published book, but I maintain my point).
So, in short, it does not help the initiated or the uninitiated. For the initiated it needs more detail for it to be useful, the uninitiated more context for it to be understood. Still a fun project, even if oversold.
[1] https://spreadsheets-are-all-you-need.ai/ [2] https://github.com/rasbt/LLMs-from-scratch
- jadengeller 15 hours agothis comment seems to be astroturfing to sell a course
