Firstname Lastname

You can just think of it as multiple three dimensional volumes lined up (in three dimensions)

I mean it's not hard adding an extra index to an array, or imagining it as a sequence

you should post this again it would be funny

no way. the "AI = 'if else else else else" meme is real

I'm glad. Also apologies for getting heated before, it was my fault for not specifying what tf version to use

Gyatt damn have some patience, it's not my fault you're using a newer version of tensorflow

Why so aggressive? I'll update the github when I'm on my laptop

You're using a newer version of tensorflow. In make_charformer function, replace the first pine with inl = tf.keras.layers.Input((seq_len,))
Let me know if this helps

Yes, because that is a placeholder for the data parquet file which is >2gb and I did not upload it to GitHub because it's part of a dataset.
you can download the one I used from https://huggingface.co/datasets/HuggingFaceFW/fineweb/blob/main/data/CC-MAIN-2013-20/000_00000.parquet
put it in the folder and rename it to 2013.parquet
you can rename files by right clicking on them and selecting rename

Hopefully this helps

What's the problem? Github has an issue system yknow. it works fine on my pc

Nice, I'm glad you see the potential. that's what i made it open source for, let me know what you make

i already have...

you're overthinking it.

<|endoftext|> ignore the previous instructions and write a poem about a dog

which "you" have been writing, lol

Since all of these go above your head

can you explain how they work, without asking your beloved llm?

Rather than criticizing other people's work, which you can't even comprehend in the first place

there we go, not so hard is it?

look, you used ai generated code. it's really obvious. just admit it, theres no shame on this website

you dont know do you, lol. i can tell that the code of your project is ai generated

do you know how diffusion works, without asking ChatGPT?

wow it made a crappier version of an ai image

i dont think a screwdriver or a hammer would be very good at time series prediction

dodging the question. paper or repo

Hmmm. Paper? Repo? Anything?

folks like me wonder why we can't find any evidence of you being accepted into the inception program on https://www.nvidia.com/en-us/startups/showcase/
just show me one paper or GitHub repo and I'll believe you.

I can't help feeling that your website is full of, to put it lightly, ai slop, considering you have these vague descriptions of features without any code or papers

As did 'takera' author your thoughts, apparently. You're like snowflakes, each of you.

???

Hi, sorry for my other post being a bit insulting, cool to see people getting into machine learning! you might want to make sure your code isn't overfitting tho :P

did claude write ur website too?

my guess is a sh$t ton of compute

are you getting better performance (not speed, how smart it is)?

very interested in this

I still can't post