InferenceX
EN
·推理图表
加载 SKU 导航器……
加载数据点元信息……

Input sequence length distribution

Output sequence length distribution

交互性随时间变化

— points

TTFT over time

— points

KV cache utilization over time

Request queue depth

Prefix cache hit rate per interval

Throughput (input & decode)

Cumulative prompt token source breakdown

Total unique input tokens over time

Unique input tokens in flight