Catalog source

github.com/MicheleCampi/inferscope

1 public CLI tools. The catalog syncs source data and can lag behind the upstream project.

Rust profiler for OpenAI-compatible LLM inference engines. Per-token timing correlated with /proc and NVML on one clock — per-device GPU metrics, NVML energy (tokens/joule), KV-cache hit-rate, per-step attribution on agentic trajectories. Apache-2.0.

Documentationgithubالمصدر بالإنجليزية

تم التحديث 2026-09-20AI & Agents