From 2f6e14a54bfca923d47a33366880a5a4c096e8c2 Mon Sep 17 00:00:00 2001
From: rnwang04 <ruonan1.wang@intel.com>
Date: Thu, 15 May 2025 09:14:59 +0000
Subject: [PATCH] fix md typo, fix code style, and update setup value error
 message

---
 README.md                            | 2 +-
 doc/README.md                        | 1 +
 doc/en/xpu.md                        | 1 -
 ktransformers/models/custom_cache.py | 2 +-
 setup.py                             | 6 +++---
 5 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/README.md b/README.md
index 172694a..cbd6e1d 100644
--- a/README.md
+++ b/README.md
@@ -23,7 +23,7 @@ Our vision for KTransformers is to serve as a flexible platform for experimentin
 
 <h2 id="Updates">🔥 Updates</h2>
 
- **May 14, 2025**: Support Intel Arc GPU ([Tutorial](./doc/en/xpu.md)).
+* **May 14, 2025**: Support Intel Arc GPU ([Tutorial](./doc/en/xpu.md)).
 
 * **Apr 29, 2025**: Support AMX-Int8、 AMX-BF16 and Qwen3MoE ([Tutorial](./doc/en/AMX.md))
 
diff --git a/doc/README.md b/doc/README.md
index 1e994aa..199b990 100644
--- a/doc/README.md
+++ b/doc/README.md
@@ -21,6 +21,7 @@ interface, RESTful APIs compliant with OpenAI and Ollama, and even a simplified
 Our vision for KTransformers is to serve as a flexible platform for experimenting with innovative LLM inference optimizations. Please let us know if you need any other features.
 
 <h2 id="Updates">🔥 Updates</h2>
+
 * **May 14, 2025**: Support Intel Arc GPU ([Tutorial](./en/xpu.md)).
 * **Apr 9, 2025**: Experimental support for LLaMA 4 models ([Tutorial](./en/llama4.md)).
 * **Apr 2, 2025**: Support Multi-concurrency. ([Tutorial](./en/balance-serve.md)).
diff --git a/doc/en/xpu.md b/doc/en/xpu.md
index 4f71ae7..7a03203 100644
--- a/doc/en/xpu.md
+++ b/doc/en/xpu.md
@@ -41,7 +41,6 @@ Install PyTorch with XPU backend support and [IPEX-LLM](https://github.com/intel
 pip install --pre --upgrade ipex-llm[xpu_2.6] --extra-index-url https://download.pytorch.org/whl/xpu
 pip uninstall torch torchvision torchaudio
 pip install torch==2.7+xpu torchvision torchaudio --index-url https://download.pytorch.org/whl/test/xpu # install torch2.7
-pip install packaging ninja cpufeature numpy
 pip uninstall intel-opencl-rt dpcpp-cpp-rt
 ```
 
diff --git a/ktransformers/models/custom_cache.py b/ktransformers/models/custom_cache.py
index 1121b4a..350af73 100644
--- a/ktransformers/models/custom_cache.py
+++ b/ktransformers/models/custom_cache.py
@@ -293,7 +293,7 @@ class KGQACache(nn.Module):
         self.v_caches = []
         
 
-    def load(self, inference_context: "sched_ext.InferenceContext"): 
+    def load(self, inference_context: "sched_ext.InferenceContext"):
         print(self.config.num_hidden_layers)
         for i in range(self.config.num_hidden_layers):
             self.k_caches.append(
diff --git a/setup.py b/setup.py
index b8f318d..0961d93 100644
--- a/setup.py
+++ b/setup.py
@@ -229,7 +229,7 @@ class VersionInfo:
         elif torch.xpu.is_available():
             backend_version = f"xpu"
         else:
-            raise ValueError("Unsupported backend: CUDA_HOME MUSA_HOME ROCM_HOME all not set.")
+            raise ValueError("Unsupported backend: CUDA_HOME MUSA_HOME ROCM_HOME all not set and XPU is not available.")
         package_version = f"{flash_version}+{backend_version}torch{torch_version}{cpu_instruct}"
         if full_version:
             return package_version
@@ -501,7 +501,7 @@ class CMakeBuild(BuildExtension):
         elif KTRANSFORMERS_BUILD_XPU:
             cmake_args += ["-DKTRANSFORMERS_USE_XPU=ON", "-DKTRANSFORMERS_USE_CUDA=OFF"]
         else:
-            raise ValueError("Unsupported backend: CUDA_HOME, MUSA_HOME, and ROCM_HOME are not set.")
+            raise ValueError("Unsupported backend: CUDA_HOME, MUSA_HOME, and ROCM_HOME are not set and XPU is not available.")
         
         cmake_args = get_cmake_abi_args(cmake_args)
         # log cmake_args
@@ -628,7 +628,7 @@ elif MUSA_HOME is not None:
 elif torch.xpu.is_available(): #XPUExtension is not available now.
     ops_module = None
 else:
-    raise ValueError("Unsupported backend: CUDA_HOME and MUSA_HOME are not set.")
+    raise ValueError("Unsupported backend: CUDA_HOME ROCM_HOME MUSA_HOME are not set and XPU is not available.")
 
 if not torch.xpu.is_available():
     ext_modules = [