Mihai

So a little bit of changes of plans. So my project got rejected cause the Demo couldn’t be a google colab page, but I never saved my my model (kinda dumb, right😁😭). So the only solution was to train it again. And I also changed some things with this occasion: Not much but modified learning rate and Input size and number of epochs (sessions of training). After all the training and modifications my AI did better in testing than the previous version:
Like the sentences now are meaningful (but not so related to each other, like in a real story), but with adjustments and training it more times I guess it will make good stories ^_____^
I also launched it on Hugging Face (find link in readme). There you can talk to my SLM (small language model)

Use: Give him a sentence without period and he will continue in a story (if it can be called story….More like random propositions, with a slightly relation between them).

Also look at release to find a portable app, kinda of a offline version.

I officially finished it I trained the AI.
Here I won’t get a lot in detail, cause it is too advanced, but I can explain some things:

1. The GPU is faster than the CPU at small tasks. So using CUDA a platform for parallel computing made by NVIDIA, I can send most of the training part to the GPU.
2. I declared the optimizer. The optimizer modifies weights in the Neuronal Network. Weights are learnable, numerical parameters that connect neurons, regulating the strength of signal transmission between them (importance, in normal language).
3. I also declared some warmups steps for my AI. Also my AI will have a higher learning rate at start and a smaller one at finish. That helps the AI to fine-tune when he already has a lot of info.

That’s all hope you learned something about AI!!!🥳😝
P.S. I also published the code on Google Colab and Github repo.

I declared my GPT style arhitecture for my NN (neuronal network).

Let me explain you how it works:
Sorry cause I couldn’t write all here cause of the character limit:
You can see it here: https://pastebin.com/vs8KZyJd

See you when the model will be trained!!!

Hi everyone! As we reach the end of this journey together, I want to sincerely thank you for following along and supporting me in #flavortown and #flavortown-help!

In this slightly long devlog, I:

Made a YouTube video so you can see if my project is worth trying.

Published the latest release of Ftown.

Updated the docs, README, and some parts of the code—including the -help text—to make it better for you!

Fixed an installation bug.

I hope you enjoyed the project and its atmosphere. Thanks for all , Flavortown!

Extra links:
https://youtu.be/FR4tUIJUT6A
https://ftown.gitbook.io/ftown-docs-1/

Ftown is presenting his newest content: 🥁🥁🥁🥁🥁(drum roll suound)
Config files for ptinting styles, tada!!!

You heard it well you can now see the outputs in 3 diffrent styles: default(old), modern, future.
You can configure it using this simple steps: 1. create a .yaml file where you put a line: table : “(the model you want)”
2. save it in ftown using: ftown config “path”
Also you can always delete it using the - -delete flag.

Ladies and gentlemens v1.0.1 stable release is here🤫😏

In this short devlog I published the latest release 😝 and I modified the docs.

📡 BREAKING FROM FTOWN 📡

After days of static, sparks flying, and near-catastrophic antenna malfunctions…
we’ve done it. The antenna is fixed. Signal restored. The airwaves are ours again.

And now… presenting the brand-new command:

ftown devlog info [ID]

Yes, you guessed right – you can read devlogs directly through the FTOWN CLI!

It’s a little slow for now, but improvements are coming in the next release.

Also check my slack channel #ftown-cli-official

!!!!This devlog was entirely made by me, no AI was used.!!!

v1.0.0 Beta Released

Version v1.0.0 Beta is officially released 🎉 🎊🎊🥳🥳🪅
You can now download the tool and use it directly from your Windows terminal.

During this session, I focused on:

• Fixing several bugs discovered during testing
• Creating a dedicated documentation page
• Completing and polishing the README with usage instructions

This release makes the project usable end-to-end for the first time.

Project Commands Added

In this session, I added core project-related commands. I’m still learning Go, so progress may be slower, but the foundation is now solid.

New commands:

• ftown project create - -title - -description
  (additional flags exist, but these are required)
• ftown project info
• ftown project update
  (uses flags to specify what you want to update, see in documentation)

For the last two commands, you can also use:

• • -byname or -n to search or update a project by name
    (note: this can be imprecise if multiple projects share the same name)

You may also notice improvements in code structure compared to earlier versions — I’m actively improving my Go skills.

Demo videos (uploads were too large for Flavortown):
https://youtu.be/ggR4gG8ST5Y
https://youtu.be/20m4DYZhGqk
https://youtu.be/jKU1LrCFFlk

Shop Commands Implemented
1h 51m logged

Added two new shop-related commands:

• ftown shop [region]
  Displays the entire shop and shows whether items are available for purchase in the specified region.

• ftown shop “[item name]” [region]
  Displays a single shop item by exact name (typos are not allowed).

These commands make browsing the Flavortown shop directly from the terminal possible.

Account Info Added

Implemented a command that retrieves and displays account information linked to the current API key.

Added running informations in readme.

API Key Management

Added commands to:

• Register an API key securely in Windows Credential Manager
• Delete the stored API key when needed

This improves security and removes the need to re-enter the key manually every time.

You can also verify your key at the moment you introduce it with the -v, - -verify flag

I added Input-Output pairs. You might wonder what that is. Let me explain: Input-Output pairs are sequences of words that the AI uses to learn what comes next. The AI doesn’t “think” like a human—it looks at all the sequences it has seen during training and learns patterns to predict the next word.

Let’s take a closer look:
a) Given the dataset:

Anna lives with her mother and her father.

Joe lives with his grandma.

Maria lives with her cousins.

b) If the AI sees the sentence:
“Joey lives”

c) It will try to predict the next word based on patterns it learned:

It has seen sequences like “lives with her mother” and “lives with his grandma.”

Based on these patterns, it predicts probable next words (“with his grandma”) rather than combining all sentences into one.

So the AI might continue the sentence like:
“Joey lives with his grandma”

Over many predictions, it can generate longer sentences that follow the patterns it has learned from the dataset.

This way, the AI uses Input-Output pairs to learn likely continuations instead of memorizing or merging every sentence. 😁

P.S.This is especially used in calculating the loss.

Added a creeper whistle, that calls a charged creeper. Craftable with 2 creeper heads, 2 lightining rods and 4 sticks.

I registered my database from HuggingFace (TinyStories) and I tokenized the splits.
Tokenizing is the main reason AI can understand words. It splits the word in smaller words, based on the frequency of the in the given dataset:
Like if we have the names:
Joe
Joey
Joseph

It takes this three tokens: “Joe” and “Joey” and “Joseph” and “Jo”

“Jo” is there cause it has a high frequency.

I added a potion named friend of the bombs, that gives you 2 sconds imunity to all explosions and it can be crafted with awkward potion and flint.

Modified the recipe and made the grenade exploadable and throwable.

Ok. So I added a recepie for the grenade. Unfortunatly it isn’t throwable and exploadable yet.

After learning java, trying several times and learning fabric and .json files I finally added a grenade item.

I worked a little more to log the ret of my minutes that weren’t logged. I rearrenged some vias to be more correct.

I finally resolved all the violation errors. Just 2 errors remained (they are totally unnecessary in this case).

I added my signature on the board and I tried to add images, but it didn’t go well.

I finished all my PCB and I added a GND filling. I also got lost of any important errors from drc.

I finished tracing the routes!!!

After a ton of work, I rearranged my pieces in a more ergonomic way (meaning a layout that makes tracing easier).

I arranged all the parts on the PCB and routed the USB-C and flash memory, although it is still a bit messy.

I changed the flash memory to a smaller one. I also arranged some of the parts on the PCB.

I added footprints to all the pieces, placed my components on the PCB, and created a rectangle for the board size.

I fisnished my schematic!!!!!

I completed the schematic configuration for the USB-C port and crystal oscillator.

I added a RP2040 microcontroller and i created some decoupling capacitors + I added labels to the GPIO pins. I also added a USB-C port. All inside kicad schematic.