Update README.md

README.md

@@ -24,16 +24,28 @@ We are continuing to train this model and will update with more evaluations and
 
 ## Deployment
 
-Deploy with vLLM (main) using the speculator as a draft model:
+Deploy with vLLM (main/nightly) using the speculator as a draft model.
 
+<details><summary>First install vllm nightly</summary>
+  
 ```bash
-vllm serve -tp 2 RedHatAI/gemma-4-31B-it-speculator.dflash --attention-backend FLASH_ATTN
+uv pip install -U vllm \
+    --torch-backend=auto \
+    --extra-index-url https://wheels.vllm.ai/nightly
+```
+
+</details>
+
+Then run:
+
+```bash
+vllm serve -tp 2 RedHatAI/gemma-4-31B-it-speculator.dflash
 ```
 
 It can also be deployed with a quantized verifier for even better speedups:
 
 ```bash
-vllm serve RedHatAI/gemma-4-31B-it-FP8-block   --tensor-parallel-size 2   --attention-backend FLASH_ATTN   --speculative-config '{
+vllm serve RedHatAI/gemma-4-31B-it-FP8-block   --tensor-parallel-size 2  --speculative-config '{
     "model": "RedHatAI/gemma-4-31B-it-speculator.dflash",
     "num_speculative_tokens": 8,
     "method": "dflash"