1 Commits

Author SHA1 Message Date
Vijay Janapa Reddi
e256ab53f9 docs(blog): 'How Much Memory Does Llama-3 70B Need?' — first blog post
SEO-optimized blog post answering the #1 question ML engineers Google.
Shows weights (140 GB FP16), KV cache (160 MB/request at 4K), and the
full serving analysis. Drives organic traffic to mlsysim.
2026-04-01 23:20:37 -04:00