📰 Dev.to · Rishabh Kharyal

Articles from Dev.to · Rishabh Kharyal · 2 articles · Updated every 3 hours · View all reads

All ⚡ AI Lessons (16886) ArXiv cs.AI Dev.to AI Dev.to · FORUM WEB Forbes Innovation Medium · Programming Medium · AI

Building a Bit-Accurate Fused QKV + RoPE Kernel for Qwen 2.5 in Triton

Dev.to · Rishabh Kharyal 🧠 Large Language Models ⚡ AI Lesson 3h ago

Building a Bit-Accurate Fused QKV + RoPE Kernel for Qwen 2.5 in Triton

How to replace 10+ PyTorch operations with a single GPU kernel while keeping the output identical to...

I Built a 4.75 Faster Qwen 2.5 Engine for a $200 GPU – Here’s How

Dev.to · Rishabh Kharyal 3h ago

I Built a 4.75 Faster Qwen 2.5 Engine for a $200 GPU – Here’s How

My RTX 3050 laptop GPU was crawling at 30 tokens per second with Qwen 2.5‑0.5B. So I tore apart the...