Keller Jordan
@kellerjordan0
CIFAR-10 fanatic
Pretraining @OpenAI OpCo LLC.
加入 March 2016
428 正在關注 16.9K 粉絲
New modded-NanoGPT optimization benchmark result: @wen_kaiyue has improved upon both the Muon and AdamW baselines, by replacing their weight decay with hyperball optimization. The new record is 3325 steps.
顯示更多
