ipex-llm 2.3.0b20250604__py3-none-manylinux2010_x86_64.whl → 2.3.0b20250606__py3-none-manylinux2010_x86_64.whl

ipex_llm/vllm/xpu/engine/engine.py

@@ -385,5 +385,5 @@ def run_mp_engine(vllm_config: VllmConfig, usage_context: UsageContext,
         engine_alive.value = False
         raise e  # noqa
 
-if os.getenv("VLLM_USE_V1"):
+if os.getenv("VLLM_USE_V1") == 1:
     IPEXLLMAsyncLLMEngine = IPEXLLMAsyncV1Engine

ipex_llm/vllm/xpu/model_convert.py

@@ -129,9 +129,15 @@ def get_load_function(low_bit):
                 if "glm-4v" in self.vllm_config.model_config.model.lower() and \
                         low_bit in ("sym_int4", "woq_int4"):
                     modules = ["dense_4h_to_h"]
+                if "phi4mm" in self.vllm_config.model_config.hf_config.model_type:
+                    modules = ["vision_encoder", "embed_tokens_extend"]
                 if low_bit == "fp16":
                     # to fix qwen2.5-vl and glm-4v
-                    modules = ["vision", "visual"]
+                    if modules is None:
+                        modules = ["vision", "visual"]
+                    else:
+                        modules.append("vision")
+                        modules.append("visual")
                 optimize_model(self.model,
                                low_bit=low_bit,
                                torch_dtype=self.vllm_config.model_config.dtype,

{ipex_llm-2.3.0b20250604.dist-info → ipex_llm-2.3.0b20250606.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ipex-llm
-Version: 2.3.0b20250604
+Version: 2.3.0b20250606
 Summary: Large Language Model Develop Toolkit
 Home-page: https://github.com/intel-analytics/ipex-llm
 Author: BigDL Authors
@@ -27,7 +27,7 @@ Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine
 Requires-Dist: torch ==2.1.2+cpu ; (platform_system == "Linux") and extra == 'all'
 Requires-Dist: torch ==2.1.2 ; (platform_system == "Windows") and extra == 'all'
 Provides-Extra: cpp
-Requires-Dist: bigdl-core-cpp ==2.7.0b20250604 ; extra == 'cpp'
+Requires-Dist: bigdl-core-cpp ==2.7.0b20250606 ; extra == 'cpp'
 Requires-Dist: setuptools ; extra == 'cpp'
 Requires-Dist: onednn-devel ==2025.0.1 ; (platform_system == "Windows") and extra == 'cpp'
 Requires-Dist: onednn ==2025.0.1 ; (platform_system == "Windows") and extra == 'cpp'
@@ -60,7 +60,7 @@ Requires-Dist: transformers ==4.40.0 ; extra == 'npu'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'npu'
 Requires-Dist: torch ==2.1.2+cpu ; (platform_system == "Linux") and extra == 'npu'
 Requires-Dist: torch ==2.1.2 ; (platform_system == "Windows") and extra == 'npu'
-Requires-Dist: bigdl-core-npu ==2.7.0b20250604 ; (platform_system == "Windows") and extra == 'npu'
+Requires-Dist: bigdl-core-npu ==2.7.0b20250606 ; (platform_system == "Windows") and extra == 'npu'
 Provides-Extra: serving
 Requires-Dist: py-cpuinfo ; extra == 'serving'
 Requires-Dist: fschat[model_worker,webui] ==0.2.36 ; extra == 'serving'
@@ -80,9 +80,9 @@ Requires-Dist: setuptools <70.0.0 ; extra == 'xpu'
 Requires-Dist: torch ==2.1.0a0 ; extra == 'xpu'
 Requires-Dist: torchvision ==0.16.0a0 ; extra == 'xpu'
 Requires-Dist: intel-extension-for-pytorch ==2.1.10+xpu ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250604 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250604 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250604 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250606 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250606 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250606 ; extra == 'xpu'
 Provides-Extra: xpu-2-1
 Requires-Dist: py-cpuinfo ; extra == 'xpu-2-1'
 Requires-Dist: protobuf ; extra == 'xpu-2-1'
@@ -97,9 +97,9 @@ Requires-Dist: setuptools <70.0.0 ; extra == 'xpu-2-1'
 Requires-Dist: torch ==2.1.0a0 ; extra == 'xpu-2-1'
 Requires-Dist: torchvision ==0.16.0a0 ; extra == 'xpu-2-1'
 Requires-Dist: intel-extension-for-pytorch ==2.1.10+xpu ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250604 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250604 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250604 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250606 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250606 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250606 ; extra == 'xpu-2-1'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-2-1'
 Requires-Dist: dpcpp-cpp-rt ==2024.0.2 ; (platform_system == "Windows") and extra == 'xpu-2-1'
 Requires-Dist: mkl-dpcpp ==2024.0.0 ; (platform_system == "Windows") and extra == 'xpu-2-1'
@@ -117,9 +117,7 @@ Requires-Dist: setuptools ; extra == 'xpu-2-6'
 Requires-Dist: torch ==2.6.0+xpu ; extra == 'xpu-2-6'
 Requires-Dist: torchvision ==0.21.0+xpu ; extra == 'xpu-2-6'
 Requires-Dist: torchaudio ==2.6.0+xpu ; extra == 'xpu-2-6'
-Requires-Dist: bigdl-core-xe-all ==2.7.0b20250604 ; extra == 'xpu-2-6'
-Requires-Dist: onednn-devel ==2025.0.1 ; extra == 'xpu-2-6'
-Requires-Dist: onednn ==2025.0.1 ; extra == 'xpu-2-6'
+Requires-Dist: bigdl-core-xe-all ==2.7.0b20250606 ; extra == 'xpu-2-6'
 Requires-Dist: dpcpp-cpp-rt ==2025.0.2 ; extra == 'xpu-2-6'
 Provides-Extra: xpu-2-6-arl
 Requires-Dist: py-cpuinfo ; extra == 'xpu-2-6-arl'
@@ -132,9 +130,7 @@ Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-2-6-arl'
 Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-2-6-arl'
 Requires-Dist: tabulate ; extra == 'xpu-2-6-arl'
 Requires-Dist: setuptools ; extra == 'xpu-2-6-arl'
-Requires-Dist: bigdl-core-xe-all ==2.7.0b20250604 ; extra == 'xpu-2-6-arl'
-Requires-Dist: onednn-devel ==2025.0.1 ; extra == 'xpu-2-6-arl'
-Requires-Dist: onednn ==2025.0.1 ; extra == 'xpu-2-6-arl'
+Requires-Dist: bigdl-core-xe-all ==2.7.0b20250606 ; extra == 'xpu-2-6-arl'
 Requires-Dist: dpcpp-cpp-rt ==2025.0.2 ; extra == 'xpu-2-6-arl'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-2-6-arl'
 Requires-Dist: torch ==2.6.0+xpu ; (platform_system == "Linux") and extra == 'xpu-2-6-arl'
@@ -143,7 +139,11 @@ Requires-Dist: torchaudio ==2.6.0+xpu ; (platform_system == "Linux") and extra =
 Requires-Dist: torch ==2.6.0.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-2-6-arl'
 Requires-Dist: torchvision ==0.21.0.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-2-6-arl'
 Requires-Dist: torchaudio ==2.6.0.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-2-6-arl'
+Requires-Dist: onednn-devel ==2025.0.1 ; (platform_system == "Windows") and extra == 'xpu-2-6-arl'
+Requires-Dist: onednn ==2025.0.1 ; (platform_system == "Windows") and extra == 'xpu-2-6-arl'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-2-6'
+Requires-Dist: onednn-devel ==2025.0.1 ; (platform_system == "Windows") and extra == 'xpu-2-6'
+Requires-Dist: onednn ==2025.0.1 ; (platform_system == "Windows") and extra == 'xpu-2-6'
 Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu'
 Requires-Dist: dpcpp-cpp-rt ==2024.0.2 ; (platform_system == "Windows") and extra == 'xpu'
 Requires-Dist: mkl-dpcpp ==2024.0.0 ; (platform_system == "Windows") and extra == 'xpu'

{ipex_llm-2.3.0b20250604.dist-info → ipex_llm-2.3.0b20250606.dist-info}/RECORD

@@ -260,16 +260,16 @@ ipex_llm/vllm/cpu/entrypoints/openai/cli_args.py,sha256=hB398yYtKauASRzevctScdbF
 ipex_llm/vllm/xpu/__init__.py,sha256=zBSG6nzrVF5QnpR6_f7kPhBFeowTE9gaZ7D5m98E7_w,585
 ipex_llm/vllm/xpu/ipex_llm_v1_wrapper.py,sha256=pd939vFomKIg9Qn2NO4u0OF6hPgvQpqcfJSxqBzcqhA,825
 ipex_llm/vllm/xpu/ipex_llm_wrapper.py,sha256=_CbhvBuf_KPnmLfngYKtJl5gPAHVsG2mWth3wSeaH3M,892
-ipex_llm/vllm/xpu/model_convert.py,sha256=oedafTsnysTi78PGYcjn1w5rnIBfBx4_mpZp2fF6z44,10093
+ipex_llm/vllm/xpu/model_convert.py,sha256=HZeTrQHMYfgXlz1b9KiKdAUZ57nLgpv6VhM5CkiSrUc,10416
 ipex_llm/vllm/xpu/engine/__init__.py,sha256=sOvwLx_Zj0jiRCGj9W3DgGTfcSU3hABYhgIQI7T6cxU,879
-ipex_llm/vllm/xpu/engine/engine.py,sha256=XAprw7VifjfnR915TZOaKcxe3QCFsVBgxzS8qOdn1yg,14462
+ipex_llm/vllm/xpu/engine/engine.py,sha256=Gzmw2TPZO0sBkwEF8Yd_9D9YrcQbFD6EEA5jG4WCezE,14467
 ipex_llm/vllm/xpu/entrypoints/openai/api_server.py,sha256=VlmS56hBHBZTIZ5Jhvb4TZN-h28O7uMn33hX8NiJXKk,45719
 ipex_llm/vllm/xpu/entrypoints/openai/cli_args.py,sha256=hB398yYtKauASRzevctScdbFIjiiSGMAe1bwEuIHrhY,10893
-ipex_llm-2.3.0b20250604.data/scripts/ipex-llm-init,sha256=fLQsT2dRL6H5bThb4GuIWotAuqoLsIxFwA-0c2qmaO8,6672
-ipex_llm-2.3.0b20250604.data/scripts/llm-chat,sha256=TdUnUmNapzuoe1c8IzrdVOQwWEg8IqsMSBRlOD3daZM,2249
-ipex_llm-2.3.0b20250604.data/scripts/llm-cli,sha256=RXGPlLElHxcKzoUxljEMBIAXbzCDysXL-Nxw-xF-7LU,2457
-ipex_llm-2.3.0b20250604.dist-info/METADATA,sha256=E9hbz-rGmyrCUOmWC7ZcofUNbDTTJRerkDK_aCPMLek,8865
-ipex_llm-2.3.0b20250604.dist-info/WHEEL,sha256=PPJcBMAZibF_2GFE9NmOJGqiaSMPiNFbJd6QaJjdA6Y,109
-ipex_llm-2.3.0b20250604.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
-ipex_llm-2.3.0b20250604.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
-ipex_llm-2.3.0b20250604.dist-info/RECORD,,
+ipex_llm-2.3.0b20250606.data/scripts/ipex-llm-init,sha256=fLQsT2dRL6H5bThb4GuIWotAuqoLsIxFwA-0c2qmaO8,6672
+ipex_llm-2.3.0b20250606.data/scripts/llm-chat,sha256=TdUnUmNapzuoe1c8IzrdVOQwWEg8IqsMSBRlOD3daZM,2249
+ipex_llm-2.3.0b20250606.data/scripts/llm-cli,sha256=RXGPlLElHxcKzoUxljEMBIAXbzCDysXL-Nxw-xF-7LU,2457
+ipex_llm-2.3.0b20250606.dist-info/METADATA,sha256=SnzYMsnRu41RAd3x_3k1wTDv3n3_4h0Pt1H7zYa86dM,9005
+ipex_llm-2.3.0b20250606.dist-info/WHEEL,sha256=PPJcBMAZibF_2GFE9NmOJGqiaSMPiNFbJd6QaJjdA6Y,109
+ipex_llm-2.3.0b20250606.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
+ipex_llm-2.3.0b20250606.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
+ipex_llm-2.3.0b20250606.dist-info/RECORD,,