Add vLLM run instructions

README.md

@@ -31,6 +31,20 @@ Gemma 4 can now be run and fine-tuned in [Unsloth Studio](https://unsloth.ai/doc
 
 See all versions of Gemma 4 (GGUF, 16-bit etc.) [in our collection](https://huggingface.co/collections/unsloth/gemma-4).
 
+## Run in vLLM
+
+```shell
+uv venv unsloth-nvfp4-env --python 3.13
+source unsloth-nvfp4-env/bin/activate
+uv pip install "vllm>=0.25.0" "flashinfer-python>=0.6.13" "nvidia-cutlass-dsl>=4.5.2" --torch-backend=auto
+```
+
+```shell
+vllm serve unsloth/gemma-4-E2B-it-NVFP4
+```
+
+Do not use the Marlin backend (around 2x slower); let vLLM auto-select the NVFP4 kernel.
+
 ---
 
 <div align="center">