Thread 16718152 - /sci/ [Archived: 468 hours ago]

Anonymous
7/7/2025, 4:28:54 AM No.16718152
nerd
nerd
md5: 3d277efe22168cc73d7cd588ff4bf828๐Ÿ”
who tf goes on 4Chan to talk abt fucking math yall r LOSERSSSSS XP
Replies: >>16718156
Anonymous
7/7/2025, 4:32:39 AM No.16718156
>>16718152 (OP)
Uhm... Rude?!

(I very much do look like this, thoughever)
Anonymous
7/7/2025, 7:26:48 AM No.16718281
dcgan_generator
dcgan_generator
md5: f7ccc945381b4c31b214732524cd4359๐Ÿ”
I just build a Deep Convolutional Generative Adverserial Network (DCGAN) to generate realistic images of faces. It will take 5 epochs and 1583 steps each epoch to train the neural network.

As my AMD GPU is not compatible with Python's machine learning packages, I have to use my GPU, taking 2.5 minutes per 50 steps. By my calculations it will take 6 hours to train the neural network, and I will be sitting here watching it the whole time to monitor for errors.

I had to learn a lot of math just to understand these models, the research paper for DCGANs was published recently in 2015. This is the crown project of my portfolio and in sha Allah these months of effort will culminate in some sort of employment that pays beyond the poverty wages I am used to doing hard manual labor.

Math skills lead to many opportunities in life. Top AI researchers at large tech companies are being offered hundreds of millions of dollars. However, I do not want to work for big tech as they uphold zionist genocide. Just a decent programming job somewhere is fine. I love mathematics and math has given me an opportunity to escape poverty.
Replies: >>16718283 >>16718286 >>16718394
Anonymous
7/7/2025, 7:30:00 AM No.16718283
Screenshot_20250707_002824_Gallery
Screenshot_20250707_002824_Gallery
md5: 1023e7625ad40de6bec1dc22a6b11c7a๐Ÿ”
>>16718281

>As my AMD GPU is not compatible with Python's machine learning packages, I have to use my CPU*

This shit will take forever. It has to perform multivariable calculus on hundreds of dimensions just to make tiny adjustments to reduce error. But it will be worth it. I have spent countless hours obsessively studying math, algorithms, and data structures just to do this.
Replies: >>16718286
bodhi
7/7/2025, 7:34:14 AM No.16718286
>>16718281
>>16718283
you have never had a crumb of pussy have you?
Replies: >>16718342
Anonymous
7/7/2025, 9:50:57 AM No.16718342
Screenshot_20250707_024834_Gallery
Screenshot_20250707_024834_Gallery
md5: 5aeba745ec5d5ac40bfcd81ee095070e๐Ÿ”
>>16718286

I have and I regret it. I wish I saved myself for marriage, speaking anonymously. Am physically fit and have no issue attracting women.

Fornication is Haram, if only I found Islam earlier in life. I wasted much of my 20s chasing empty pleasure, intoxication and fornication.

Now I can actually use my brain and apply my potential as a human being. If I could go back in time, I would have learned advanced math in my early 20s.

About 2.5 hours in, almost halfway done training the model. I think it will be done by morning light.
Replies: >>16718347
bodhi
7/7/2025, 10:01:55 AM No.16718347
>>16718342
I just joshing you. Do what you love
Anonymous
7/7/2025, 10:53:48 AM No.16718394
>>16718281
> I do not want to work for big tech as they uphold zionist genocide

and in rebellion youโ€™re gonna let the jews CONTINUE to make the money you could be making? thank god moral compass retards like you leave free lunch for us on the table
Replies: >>16718440
Anonymous
7/7/2025, 12:39:51 PM No.16718440
Screenshot_20250707_053459_Gallery
Screenshot_20250707_053459_Gallery
md5: 2cbadbb1881bee7172c8e04b80bfe98d๐Ÿ”
>>16718394

You call me a "retard" because I believe in God and disagree with mass murdering children and making a profit on wars? Making money is easy if you have no morals. No math skills required. Pornographers, landlords, drug dealers, war profiteers.

Those people make more money than any sort of engineer or scientist. I learn math so I can succeed WITHOUT selling my soul. You call it a "free lunch" to have no morals and whore yourself to the highest bidder? What's the point in all this hard work if you are a degen anyways? You sell yourself short to Satan really. Very unwise.
Replies: >>16718442
Anonymous
7/7/2025, 12:44:28 PM No.16718442
quote-a-year-spent-in-artificial-intelligence-is-enough-to-make-one-believe-in-god-alan-perlis-22-90-22(2)
>>16718440

And zionism is not Judaism. Many Jews in science, Alan Perlis is one of my inspirations. Also a God believing man. These people actually care about humanity and making a difference, true scientists, not sellout #927263.
Replies: >>16718457
Anonymous
7/7/2025, 1:15:38 PM No.16718457
Screenshot_20250707_061159_Telegram
Screenshot_20250707_061159_Telegram
md5: cf54b6178d4774d303a978f47f744ffc๐Ÿ”
>>16718442

Six hours of calculation and it threw an error when plotting the results because I forgot a single equals sign for single variable. It was so dumb I laughed out loud. At least my loss chart was plotted, looks like a success. Have to restart the whole thing to get the full results. Math is actually a lot of fun.
Replies: >>16718743
Anonymous
7/7/2025, 8:41:10 PM No.16718743
Screenshot_20250707_133711_Telegram
Screenshot_20250707_133711_Telegram
md5: 0cd4f350b32e978ea2fcbf4b3a40e8b3๐Ÿ”
>>16718457

Finally managed to plot the results. Wow, pretty horrifying. Works as a proof of concept at least. Wish I had access to more computing power. Even a GPU or two that is compatible.
Replies: >>16718777
Anonymous
7/7/2025, 9:10:14 PM No.16718777
Screenshot_20250707_140513_Firefox
Screenshot_20250707_140513_Firefox
md5: ee8c891ea3448488fd430453a347a63b๐Ÿ”
>>16718743

https://arxiv.org/pdf/1511.06434

Here is the paper the model is based off of. The same quality results were successfully reproduced. With more computing power, more impressive results could be made with larger images and vectors with many more epochs. But overall it was an experimental success. Science!
Replies: >>16719543
Anonymous
7/8/2025, 4:43:03 PM No.16719543
>>16718777
Hooray!!