DDraceNetwork - showroom

I.K.U

I LOVE GETTING LAG SPIKES

looks like you're getting spiked as well

The great russian invasion of 2024

1

00:17

3k players tho

00:17

00:18

@deen it's so cool to see the game growing, couldn't expect this in 2017

Astramast

The great russian invasion of 2024

согласен

triki

согласен

alright ?

00:34

idk i don't speak russian mate

triki

согласен

Astramast

The great russian invasion of 2024

what happened here tho?

the game released on steam

BratOK

what happened here tho?

New region unlocked in the dlc's

2

BratOK

what happened here tho?

https://ddnet.org/news/ddnet-year-2021-in-review/

2

Thanks!

1

11:03

solo xd

Blender

Click to see attachment 🖼️

why not use deepfly instead of switching accounts

Astramast

The great russian invasion of 2024

I wonder what youtuber played ddnet. Or why it is becoming more popular.

CovER

I wonder what youtuber played ddnet. Or why it is becoming more popular.

it got reccomended on tiktok

1

1.6M

my brother told me for this game lol, idk where he got it

user56173

why not use deepfly instead of switching accounts

deepfly release hook

Blender

deepfly release hook

rehooks

11:29

doesnt just let go

and i dont use deepfly

if you look up itll rehook up

11:29

oh

ig its possible with deepfly but i already got used to doing like that

Use w?

-

11:31

bind mouse2 "+hook; +toggle cl_Dummy_hammer 1 0"

2

2

11:31

hook=dummyhammer

11:33

@MrWiskis ???

user56173

rehooks

rehooks only where the dummy is aiming

Cammodude

rehooks only where the dummy is aiming

thats why i said to look up

Blender

@MrWiskis ???

"comfortable" bind...

yeah i just have to press 1 button instead of 2

And can't do easy trick instead

i do have deepfly on buttons far away for special things

11:37

i rarely use it

Me too

11:38

But mouse2 bind... pepeW

u racist

1

11:39

against mouse2 bind

11:40

u use w?

1

11:40

i got used to mouse from start

If this how did u call that. Yeah, I'm. Anyway it's u choice

i have dummy control shit on w

mouse2 bind haters

1

xd

11:45

i changed it because it was easier to do pseudo aleds

11:45

since you dont have to switch between pistol and hammer

1

manikoo

mouse2 bind haters

Yes I am

(edited)

Blender

since you dont have to switch between pistol and hammer

just like on w

mouse>w

instead of w, you can put a more important bind

jump for example

1

jump for L

many people use W to jump :-:

i use space

12:05

on w dummy jump)

i have pause on w

why not q

q is team_say back

zoom on mouse, weapon switching on w and s, every other bind is wrong

BratOK

what happened here tho?

steam be like : Oh hello there

Cammodude

zoom on mouse, weapon switching on w and s, every other bind is wrong

mousewheel?

meloƞ

jump for example

i have dummy jump on it

Blender

mousewheel?

yeah

imagine mouse1 zoomout mouse2 zoomin...

Cammodude

zoom on mouse, weapon switching on w and s, every other bind is wrong

mousewheel = zoom, mouse4 is dummy hammer, mouse5 is dummy copy

Already use zoom control on mousewheel

y for /spec, w for /pause also i have a lot of ctrl+ binds

mousewhell weapons mouse4 dummy copy moves mouse5 kill

1

meloƞ

mousewheel = zoom, mouse4 is dummy hammer, mouse5 is dummy copy

4 is resetonswitch toggle, 5 is deepfly toggle

xd

But Ctrl+mousewheel

manikoo

mousewhell weapons mouse4 dummy copy moves mouse5 kill

ctrl as kill troll

my teleport and reset binds for practice are ctrl+q ctrl+w ctrl+e

12:07

dummyreset is ctrl+s

i was using zoom for mousewheel before but i raged and broke my mousewheel somehow

lol

12:07

oh wait we're in showroom, go move this to #general !

1

cringe

12:08

bind on kill is bad

12:08

so many runs ruined

Ya?

12:08

Only?

why showroom isnt media only?

Teeworlds/DDNet related media only

12:13

other stuff goes into other channels

1

Neural network training to beat not that simple gores map. Next step have to be completing Aip-Gores. (edited)

20

7

2

CovER

Neural network training to beat not that simple gores map. Next step have to be completing Aip-Gores. (edited)

when will it gain sentience

1

Are you gonna do some crazy video like "I trainer AI for a 100 hours on a kog map"? I would honestly watch something like that if you still have the evolution

1

user56173

when will it gain sentience

Maybe someday

Devinci

Are you gonna do some crazy video like "I trainer AI for a 100 hours on a kog map"? I would honestly watch something like that if you still have the evolution

I thought about it, but probably no

rip

the tee river

1

guys drifting

1

14:07

who stole my points

14:08

fess up

Tsumugi

who stole my points

SPOILER

2

Tsumugi

who stole my points

im a robber

CovER

Neural network training to beat not that simple gores map. Next step have to be completing Aip-Gores. (edited)

Where phd thesis?

4

15:17

We have the perfect data set to train a DDNet-playing AI: 10 TB of compressed player data as teehistorian files

deen

We have the perfect data set to train a DDNet-playing AI: 10 TB of compressed player data as teehistorian files

Actually I'm new to AI, but if I understand how to make offline training and stuff like that it would be really interesting if I could use that dataset.

15:25

I think I will publish the source code so you can train it using that dataset without giving it to me

undertale muffet skin (edited)

Genosse

Click to see attachment 🖼️

veri cool nice

random skill manikoo best movements

manikoo

random skill manikoo best movements

gay

no

21:05

why?

21:05

CovER

Actually I'm new to AI, but if I understand how to make offline training and stuff like that it would be really interesting if I could use that dataset.

Use it to detect cheaters :)

SPYRES

Use it to detect cheaters :)

how do u even imagine this?

Tsumugi

how do u even imagine this?

Well I am not sure what data teehistorian has. I suspect it has the right level of information for someone to use deep learning to identify non natural movements. That's the idea in a nutshell. While they have "cheat detection" mechanisms already, they can enhance that with deep learning.

22:34

I'm sure there are white papers online that you can find that reference something similar. Google Scholar is your best friend

https://scholar.google.com/

SPYRES

Well I am not sure what data teehistorian has. I suspect it has the right level of information for someone to use deep learning to identify non natural movements. That's the idea in a nutshell. While they have "cheat detection" mechanisms already, they can enhance that with deep learning.

Usually such systems put a lot of load on the processor; even now, with simply a large number of people on one server, there are problems because of this