diff --git a/Dockerfile b/Dockerfile index 4525358..646f6e1 100644 --- a/Dockerfile +++ b/Dockerfile @@ -2,7 +2,7 @@ # Extends the vLLM dream-build container with our custom DeepGEMM kernel # and DeepSeek V4 patch. -FROM atl.vultrcr.io/vllm/vllm-with-lmcache:dream-build +FROM atl.vultrcr.com/vllm/vllm-with-lmcache:dream-build # Install build essentials RUN apt-get update && apt-get install -y git screen cmake && rm -rf /var/lib/apt/lists/* diff --git a/README.md b/README.md index 57ea183..706b367 100644 --- a/README.md +++ b/README.md @@ -15,6 +15,7 @@ Full NVFP4 quantization of DeepSeek V4 Pro and vLLM serving on 8× NVIDIA B200 G | Profile/Warmup Run | ✅ Passes | | API Server | ✅ Running on port 8000 | | Output Quality | 🔧 Garbled — likely remaining dequant/scale bug | +| NVFP4 Mega MoE Kernel | 🔧 Built, pending compile & integration | ## B200 Node