Did I miss something? https://github.com/NVlabs/Fast-dLLM/blob/main/llada/chat.py

That’s inference code, but where is the high perf web server?