Andrej Karpathy
Building a kind of JARVIS @ OреոΑӏ. Previously Director of AI @ Tesla, CS231n, PhD @ Stanford. I like to train large deep neural nets 🧠🤖💥
عرض في 𝕏سلاسل التغريدات
@ch402 Love the idea for Software 3.0 😂. Programming moving from curating datasets to curating prompts to make the meta learner "get" the task it's supposed to be doing. LOL 🤣👌
One of my favorite days of the year is the GTC Keynote day, nerding out over (some big new X)FLOPS of tensor compute capability; Today the big news is the new A100 and its DGX, ann...
ResNet-50 on ImageNet now (allegedly) down to 224sec (3.7min) https://t.co/3Z77Edfj0u using 2176 V100s. Increasing batch size schedule, LARS, 5 epoch LR warmup, synch BN without mo...