LLM Inference on AMD MI300X with ROCm 6: vLLM Support and Benchmarks

本文以日语发表。中文摘要如下：

LLM Inference on AMD MI300X with ROCm 6: vLLM Support and Benchmarks—Running LLM inference on AMD MI300X with ROCm 6 and vLLM: support status, real Llama/Qwen throughput numbers vs H100 and Japan-region availability.