DeepSeek V4 Flash 284B (13B active)
on 2× NVIDIA DGX Spark · vLLM · 1,048,576 ctx
Summary
User reports DeepSeek-V4-Flash-0731 (304B MoE) at 82 tok/s decode and about 1400 tok/s prefill on 2x DGX Spark, each with 128GB unified memory.
Setup is vLLM 0.26.1rc1 with TP=2, 1M context and FP8.
User asks for advice on reducing VRAM and OS RAM usage.