[Feature] Speculative decoding support lookahead (#9873)

Co-authored-by: a4zhangfei <a4zhangfei@qq.com>
Co-authored-by: Qiaolin-Yu <liin1211@outlook.com>

python/sglang/bench_serving.py

@@ -1110,7 +1110,8 @@ def sample_sharegpt_requests(
                 add_generation_prompt=True,
                 tokenize=False,
             )
-            prompt = prompt.replace(tokenizer.bos_token, "")
+            if tokenizer.bos_token:
+                prompt = prompt.replace(tokenizer.bos_token, "")
 
         prompt_token_ids = tokenizer.encode(prompt)
         completion = dataset[i][1]