[Doc] Explicitly state that PP isn't compatible with speculative decoding yet (#10975)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/model_executor/models/exaone.py

@@ -473,10 +473,11 @@ class ExaoneForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
             self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
                                                     config.vocab_size,
                                                     logit_scale)
-            self.sampler = get_sampler()
         else:
             self.lm_head = PPMissingLayer()
 
+        self.sampler = get_sampler()
+
         self.make_empty_intermediate_tensors = (
             self.transformer.make_empty_intermediate_tensors)