popopanda

6 ms·

1. 1.▲
   by popopanda 29d ago

   Yeah, tokens/s varies a lot based on workload. However, I’ve calibrated the estimator against public benchmarks, so it stays within a 30% error margin! I'll definitely explore how to estimate vision models next!
2. 2.▲
   Show HN: LLM Inference Calculator – Estimate VRAM, Latency, and Throughput(llm-inference-calculator-delta.vercel.app)

   6 points by popopanda 29d ago|4 comments
3. 3.▲
   Show HN: Minimal LLM Post-Training Experiments on an 8GB GPU (SFT, DPO, GRPO)(github.com)

   21 points by popopanda 2mo ago|0 comments