Tagged with

2 articles found

3000 Tokens Per Second? Xiaomi’s MiMo V2.5 Just Rewrote the Physics of Inference

Xiaomi’s MiMo V2.5 hits 3000 tps with a 1-trillion-parameter model using a radical FP4 quantization and a ‘block-diffusion’ drafter. Here’s the tech that made it happen and the catch.

#Deep Learning#LLM Inference#Open Source...

artificial intelligence

Kimi Just Made Residual Connections Obsolete: The 10-Year Assumption That Crumbled Overnight

Moonshot AI’s Attention Residuals architecture replaces decade-old residual connections with selective depth-wise attention, delivering 1.25x compute efficiency and breaking the PreNorm dilution bottleneck that has plagued deep transformers.

#artificial intelligence#Deep Learning#machine learning...