All courses
AIAI path · course 333 of 422
Tensor Cores and Low-Precision Math
Advanced · 3 lessons · 0 complete
Floating-point formats (FP32, BF16, FP16, FP8, FP4), why lower precision is faster, matrix-multiply units, mixed-precision training, quantisation for inference and the accuracy trade-offs.
Multichain Builders Secure checkoutSubscription
Full access subscription
$3/ month
≈ KES 390 · 30 days, no auto-renewal
Enter your Safaricom number. Nothing is charged until you approve the prompt with your M-Pesa PIN.