Update README.md

README.md

@@ -76,6 +76,8 @@ Average acceptance length (τ) measured across MT-bench, HumanEval, and GSM8K wi
 
 *Measured at temperature = 1 with K = 7*
 
+> **Note:** Earlier vLLM versions sampled draft tokens greedily regardless of temperature, which underestimated acceptance rates at temperature > 0. Stochastic draft sampling was introduced in **v0.18.0**, and from **v0.21.0** it can be enabled via `speculative_config` using `rejection_sample_method` and `draft_sample_method`. The acceptance metrics reported above were measured under standard rejection sampling and are reproducible with the configuration below.
+
 ## Usage with vLLM
 ```python
 from vllm import LLM, SamplingParams
@@ -86,6 +88,8 @@ llm = LLM(
         "method": "eagle3",
         "model": "nebius/EAGLE3-Llama-3.3-70B-Instruct",
         "num_speculative_tokens": 6,
+        "rejection_sample_method": "standard",
+        "draft_sample_method": "gumbel",
     },
 )
 
@@ -93,8 +97,6 @@ sampling_params = SamplingParams(temperature=0.7)
 outputs = llm.generate(["Explain speculative decoding in simple terms."], sampling_params)
 ```
 
-> **Note**: The current vLLM implementation samples draft tokens greedily regardless of temperature settings, which can underestimate acceptance rates at temperature > 0. A community fix is under development (see [vllm-project/vllm#20459](https://github.com/vllm-project/vllm/pull/20459)). The acceptance metrics reported above were measured with proper rejection sampling.
-
 ## License
 
 [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)