The lookahead speculative decoding failed when applied to gpt2 0.1b.

[JeRainXiong]

Hi,
My environment is A10 22G, trtllm v0.16.0, and CUDA 12.6. When I use lookahead speculative decoding on gpt2 0.1B, I encounter this error. How can I fix it？

[TensorRT-LLM][ERROR] Assertion failed: No available XQA kernels are found for speculative decoding mode.
（/target/trtllm_1118/TensorRT-LLM/cpp/tensorrt 1lm/plugins/gptAttentionCommon/gptAttentionCommon.cpp:1745）

[JeRainXiong]

I solve this problem by using TensorRT-LLM version: 0.18.0.dev2025031100