[Fix] Keep Kimi benchmark tokenizer offline

experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/run_moe_backend_prefill.sh

@@ -265,6 +265,7 @@ run_one_bench() {
     -e PYTHONUNBUFFERED=1 --entrypoint python3 "$CANDIDATE_IMAGE" \
     -m "$BENCH_MODULE" \
     --backend sglang --host "$HEAD_HOST" --port "$PORT" \
+    --tokenizer "$MODEL_PATH" \
     "${dataset_args[@]}" \
     --random-input-len "$INPUT_LEN" --random-output-len "$OUTPUT_LEN" \
     --random-range-ratio 1.0 --num-prompts "$NUM_PROMPTS" \