test: 回退Docker context到26e6cb40完全一致——验证竞赛平台build

Dockerfile/qwen3_6_scripts/ex_engine/computility-run.yaml 全部
还原到26e6cb40的精确内容。删除所有26e6cb40不存在的新增文件
(prebuilt/*.so, wheels/*.whl, vendor_overrides/, 新增.cu/.sh等)。

目的：确认26e6cb40的文件内容在当前git状态下仍能通过竞赛平台build。
如果通过，说明问题在新增文件中；如果不通过，说明问题在git仓库层面。

ex_engine/csrc/ilu/layer_attention.h

@@ -1,82 +0,0 @@
-/* Copyright 2025 The xLLM Authors. All Rights Reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    https://github.com/jd-opensource/xllm/blob/main/LICENSE
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-==============================================================================*/
-
-#pragma once
-
-#include <torch/torch.h>
-
-#include <tuple>
-
-#include "framework/kv_cache/kv_cache.h"
-#include "framework/model/model_input_params.h"
-#include "layers/common/attention_metadata.h"
-
-namespace xllm {
-namespace layer {
-class AttentionImpl : public torch::nn::Module {
- public:
-  AttentionImpl() = default;
-
-  AttentionImpl(int64_t num_heads,
-                int64_t head_size,
-                float scale,
-                int64_t num_kv_heads,
-                int64_t sliding_window);
-  AttentionImpl(int64_t num_heads,
-                int64_t head_size,
-                int64_t num_kv_heads,
-                int64_t v_head_dim,
-                int64_t sliding_window,
-                float scale,
-                bool use_fused_mla_qkv,
-                bool enable_lighting_indexer,
-                bool enable_mla);
-
-  std::tuple<torch::Tensor, c10::optional<torch::Tensor>> forward(
-      const AttentionMetadata& attn_metadata,
-      torch::Tensor& query,
-      torch::Tensor& key,
-      torch::Tensor& value,
-      KVCache& kv_cache);
-
-  void prefill_forward(torch::Tensor& query,
-                       torch::Tensor& key,
-                       torch::Tensor& value,
-                       torch::Tensor& output,
-                       const torch::Tensor& k_cache,
-                       const c10::optional<torch::Tensor>& v_cache,
-                       const AttentionMetadata& attn_metadata);
-
-  void decoder_forward(torch::Tensor& query,
-                       torch::Tensor& output,
-                       const torch::Tensor& k_cache,
-                       const c10::optional<torch::Tensor>& v_cache,
-                       const AttentionMetadata& attn_metadata);
-
- private:
-  int64_t num_heads_;
-  int64_t head_size_;
-  float scale_;
-  int64_t num_kv_heads_;
-  int64_t v_head_dim_;
-  bool use_fused_mla_qkv_;
-  bool enable_lighting_indexer_;
-  bool enable_mla_;
-  int64_t sliding_window_;
-};
-TORCH_MODULE(Attention);
-
-}  // namespace layer
-}  // namespace xllm