#optimization

0

🔗

Original post on piaille.fr (piaille.fr)

Since last release of Quicktake for Apple II, I've added: - thumbnail support for Quicktake 200 - JPEG YH1V1 decoding - RADC decoding of 768x512 images Compared to the last release, these three things take -3584 bytes (minus 3.5kB) on disk, and -1551 bytes (minus 1.5kB) in memory […]

4 comments · 13h ago · by Colin McMillen #optimization#retrocomputing

0

🔗

Everyone Should Know SIMD (mitchellh.com)

“Everyone Should Know SIMD”, Mitchell Hashimoto (mitchellh.com/writing/ever...). On HN: news.ycombinator.com/item?id=4901... On Lobsters: lobste.rs/s/gpqa52/eve...

0 comments · 15h ago · by Ranjit Mathew #intrinsics#optimization#performance

0

Using a Transformer Model: From Training to Inference - MachineLearningMastery.com (machinelearningmastery.com)

Inference isn't just training minus backpropagation. Prefill once, decode many times. KV cache saves memory but consumes it. Learn how to optimize your model for efficient inference! 🔑🧠

0 comments · 1d ago · by sergiocuellar.bsky.social #inference#optimization#transformer