Oner Yalcin PRO

oneryalcin

1 2 34

AI & ML interests

None yet

Recent Activity

liked a model 3 days ago

FrontiersMind/Lumma-fev-0.6b

liked a model 4 days ago

muose/Muose-50M-Decision

liked a model 7 days ago

convaiinnovations/laya

View all activity

Organizations

None yet

liked a model 3 days ago

FrontiersMind/Lumma-fev-0.6b

Text Classification • 0.6B • Updated about 10 hours ago • 1k • 55

liked a model 4 days ago

muose/Muose-50M-Decision

50.8M • Updated 5 days ago • 60 • 8

liked a model 7 days ago

convaiinnovations/laya

Text Classification • 0.4B • Updated 2 days ago • 3.71k

liked a model 10 days ago

harshatheg/Qwen-2.5-1B-RLCD

Text Generation • Updated 10 days ago • 569

reacted to anakin87's post with ❤️ 10 days ago

Post

3458

I made a 1.1M ModernBERT encoder play Doom in real time on a CPU

Some time ago, VAGO Solutions released SauerkrautLM-Doom-MultiVec-1.3M, a tiny model trained to play Doom Defend the Center scenario from 31k human gameplay examples.

My first thought: cool! I love both Doom and Small Language Models.

Then another idea: I bet I can do better :-)

What I did?
- evaluated the original model and found it's better than reported
- changed a bit the architecture
- generated SFT data with a scripted oracle
- SFT + PPO refinement on consumer hardware

Got a smaller, faster and killer model
Can even fit a floppy with int8 quantization 💾

Watch it play/read the article: anakin87/tiny-doom-defender