ipex-llm 2.3.0b20250421__py3-none-win_amd64.whl → 2.3.0b20250422__py3-none-win_amd64.whl

ipex_llm/transformers/model.py

@@ -775,12 +775,15 @@ class _BaseAutoModelClass:
                                                                 model)
             torch.distributed.barrier()
 
-        # add lookup_generate to loaded model
-        from .lookup import lookup_generate
-        import types
-        model.lookup_generate = types.MethodType(lookup_generate, model)
-        if model.config.model_type == "minicpmv" and hasattr(model, 'llm'):
-            model.llm.lookup_generate = types.MethodType(lookup_generate, model.llm)
+        try:
+            # add lookup_generate to loaded model
+            from .lookup import lookup_generate
+            import types
+            model.lookup_generate = types.MethodType(lookup_generate, model)
+            if model.config.model_type == "minicpmv" and hasattr(model, 'llm'):
+                model.llm.lookup_generate = types.MethodType(lookup_generate, model.llm)
+        except ImportError as e:
+            pass
 
         return model
 

ipex_llm/transformers/models/deepseek.py

@@ -32,7 +32,7 @@ from ipex_llm.utils.common.log4Error import invalidInputError
 from ipex_llm.transformers.kv import DynamicNormalCache
 from ipex_llm.transformers.models.common import padding_mla_v_hd_base
 from ipex_llm.transformers.models.common import scaled_dot_product_attention
-from ipex_llm.transformers.models.utils import rotate_half
+from ipex_llm.transformers.models.utils import rotate_half, use_fuse_moe
 
 
 def padding_mla_v_hd(module: torch.nn.Module):
@@ -291,11 +291,8 @@ def fuse_gate_forward(self, x: torch.Tensor):
 
 
 def moe_infer_decode(self, x: torch.Tensor, topk_ids: torch.Tensor, topk_weight: torch.Tensor):
-    if (
-        x.device.type == "xpu"
-        and x.dtype in [torch.float, torch.half]
-        and self.experts[0].down_proj.qtype == 2
-    ):
+    qtype = self.experts[0].down_proj.qtype
+    if use_fuse_moe(x, qtype):
         if getattr(self, "gates", None) is None:
             gate_addrs = [expert.gate_proj.weight.data_ptr() for expert in self.experts]
             up_addrs = [expert.up_proj.weight.data_ptr() for expert in self.experts]
@@ -310,7 +307,7 @@ def moe_infer_decode(self, x: torch.Tensor, topk_ids: torch.Tensor, topk_weight:
         import xe_linear
         final_out = xe_linear.moe_forward_vec(
             x, topk_ids, topk_weight, self.gates, self.ups, self.downs,
-            x.size(-1), self.experts[0].intermediate_size, 2
+            x.size(-1), self.experts[0].intermediate_size, qtype
         )
     else:
         idxs = topk_ids.flatten().tolist()

ipex_llm/transformers/models/utils.py

@@ -394,3 +394,11 @@ def make_cache_contiguous_inplaced(cos: torch.Tensor, sin: torch.Tensor):
         new_sin = sin.contiguous()
         cos.set_(new_cos)
         sin.set_(new_sin)
+
+
+def use_fuse_moe(hidden_states: torch.Tensor, qtype: int):
+    return (
+        hidden_states.device.type == "xpu"
+        and hidden_states.dtype in [torch.float, torch.half]
+        and qtype in [ggml_tensor_qtype["sym_int4"], ggml_tensor_qtype["woq_int4"]]
+    )

{ipex_llm-2.3.0b20250421.dist-info → ipex_llm-2.3.0b20250422.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ipex-llm
-Version: 2.3.0b20250421
+Version: 2.3.0b20250422
 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 ==2.7.0b20250421 ; extra == 'cpp'
+Requires-Dist: bigdl-core-cpp ==2.7.0b20250422 ; 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 ==2.7.0b20250421 ; (platform_system == "Windows") and extra == 'npu'
+Requires-Dist: bigdl-core-npu ==2.7.0b20250422 ; (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 ==2.7.0b20250421 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250421 ; extra == 'xpu'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250421 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250422 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250422 ; extra == 'xpu'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250422 ; 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 ==2.7.0b20250421 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250421 ; extra == 'xpu-2-1'
-Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250421 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-21 ==2.7.0b20250422 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-batch-21 ==2.7.0b20250422 ; extra == 'xpu-2-1'
+Requires-Dist: bigdl-core-xe-addons-21 ==2.7.0b20250422 ; 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'
@@ -117,7 +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 ==2.7.0b20250421 ; extra == 'xpu-2-6'
+Requires-Dist: bigdl-core-xe-all ==2.7.0b20250422 ; 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: dpcpp-cpp-rt ==2025.0.2 ; extra == 'xpu-2-6'
@@ -132,7 +132,7 @@ 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 ==2.7.0b20250421 ; extra == 'xpu-2-6-arl'
+Requires-Dist: bigdl-core-xe-all ==2.7.0b20250422 ; 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: dpcpp-cpp-rt ==2025.0.2 ; extra == 'xpu-2-6-arl'
@@ -144,75 +144,6 @@ Requires-Dist: torch ==2.6.0.post0+xpu ; (platform_system == "Windows") and extr
 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: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-2-6'
-Provides-Extra: xpu-arc
-Requires-Dist: py-cpuinfo ; extra == 'xpu-arc'
-Requires-Dist: protobuf ; extra == 'xpu-arc'
-Requires-Dist: mpmath ==1.3.0 ; extra == 'xpu-arc'
-Requires-Dist: numpy ==1.26.4 ; extra == 'xpu-arc'
-Requires-Dist: transformers ==4.37.0 ; extra == 'xpu-arc'
-Requires-Dist: sentencepiece ; extra == 'xpu-arc'
-Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-arc'
-Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-arc'
-Requires-Dist: tabulate ; extra == 'xpu-arc'
-Requires-Dist: setuptools ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-23 ==2.7.0b20250421 ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.7.0b20250421 ; extra == 'xpu-arc'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.7.0b20250421 ; extra == 'xpu-arc'
-Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-arc'
-Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arc'
-Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arc'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110+xpu ; (platform_system == "Linux") and extra == 'xpu-arc'
-Requires-Dist: torch ==2.3.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-arc'
-Requires-Dist: torchvision ==0.18.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-arc'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-arc'
-Requires-Dist: onednn-devel ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-arc'
-Requires-Dist: onednn ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-arc'
-Provides-Extra: xpu-arl
-Requires-Dist: py-cpuinfo ; extra == 'xpu-arl'
-Requires-Dist: protobuf ; extra == 'xpu-arl'
-Requires-Dist: mpmath ==1.3.0 ; extra == 'xpu-arl'
-Requires-Dist: numpy ==1.26.4 ; extra == 'xpu-arl'
-Requires-Dist: transformers ==4.37.0 ; extra == 'xpu-arl'
-Requires-Dist: sentencepiece ; extra == 'xpu-arl'
-Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-arl'
-Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-arl'
-Requires-Dist: tabulate ; extra == 'xpu-arl'
-Requires-Dist: setuptools ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-23 ==2.7.0b20250421 ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.7.0b20250421 ; extra == 'xpu-arl'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.7.0b20250421 ; extra == 'xpu-arl'
-Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-arl'
-Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arl'
-Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-arl'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110+xpu ; (platform_system == "Linux") and extra == 'xpu-arl'
-Requires-Dist: torch ==2.3.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-arl'
-Requires-Dist: torchvision ==0.18.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-arl'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-arl'
-Requires-Dist: onednn-devel ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-arl'
-Requires-Dist: onednn ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-arl'
-Provides-Extra: xpu-lnl
-Requires-Dist: py-cpuinfo ; extra == 'xpu-lnl'
-Requires-Dist: protobuf ; extra == 'xpu-lnl'
-Requires-Dist: mpmath ==1.3.0 ; extra == 'xpu-lnl'
-Requires-Dist: numpy ==1.26.4 ; extra == 'xpu-lnl'
-Requires-Dist: transformers ==4.37.0 ; extra == 'xpu-lnl'
-Requires-Dist: sentencepiece ; extra == 'xpu-lnl'
-Requires-Dist: tokenizers ==0.15.2 ; extra == 'xpu-lnl'
-Requires-Dist: accelerate ==0.23.0 ; extra == 'xpu-lnl'
-Requires-Dist: tabulate ; extra == 'xpu-lnl'
-Requires-Dist: setuptools ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-23 ==2.7.0b20250421 ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-batch-23 ==2.7.0b20250421 ; extra == 'xpu-lnl'
-Requires-Dist: bigdl-core-xe-addons-23 ==2.7.0b20250421 ; extra == 'xpu-lnl'
-Requires-Dist: intel-openmp ; (platform_machine == "x86_64" or platform_machine == "AMD64") and extra == 'xpu-lnl'
-Requires-Dist: torch ==2.3.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-lnl'
-Requires-Dist: torchvision ==0.18.1+cxx11.abi ; (platform_system == "Linux") and extra == 'xpu-lnl'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110+xpu ; (platform_system == "Linux") and extra == 'xpu-lnl'
-Requires-Dist: torch ==2.3.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-lnl'
-Requires-Dist: torchvision ==0.18.1.post0+cxx11.abi ; (platform_system == "Windows") and extra == 'xpu-lnl'
-Requires-Dist: intel-extension-for-pytorch ==2.3.110.post0+xpu ; (platform_system == "Windows") and extra == 'xpu-lnl'
-Requires-Dist: onednn-devel ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-lnl'
-Requires-Dist: onednn ==2024.1.1 ; (platform_system == "Windows") and extra == 'xpu-lnl'
 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-2.3.0b20250421.dist-info → ipex_llm-2.3.0b20250422.dist-info}/RECORD

@@ -41,35 +41,35 @@ ipex_llm/langchain/llms/transformerspipelinellm.py,sha256=vm522YPPwWxxAPVvQBtxRf
 ipex_llm/langchain/vllm/__init__.py,sha256=T-EbRT6GJ_8RCu-iLmSzcftOimXSPQf2d5X72AUAy2Y,874
 ipex_llm/langchain/vllm/vllm.py,sha256=6dxc-ZISZQrJilEa_HA827l75Dv9rcHpY_G6FdJ8BVs,7793
 ipex_llm/libs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-ipex_llm/libs/bloom-api.dll,sha256=2mqbLuDXjsPYhAQqdha6XexjWUOr6HlkRKtIsX0cdAI,36352
-ipex_llm/libs/bloom.dll,sha256=z-BdZ2JWQDPo4aMW4HqTShQNvn-bhUKqGlaueZ_cNlA,507904
-ipex_llm/libs/gptneox-api.dll,sha256=laQYdV3YwZzpxMu17A0yhoEB928tIvqeXxKzZyRC7H0,24576
-ipex_llm/libs/gptneox.dll,sha256=0LSiXx7vIzpc5CuqGHZtu_kwtKvwZuj_Khi1QFbh3T0,568320
-ipex_llm/libs/libbloom_avx.dll,sha256=7UyUgVKa89KrTKyWOZowBPpR1OceermvNigXj4j2VnE,536576
-ipex_llm/libs/libbloom_vnni.dll,sha256=YTRD_1eUleJ0G0b-kh_Amd7cvZrqDoBirjGw6XS1_hI,508416
-ipex_llm/libs/libgptneox_avx.dll,sha256=G1o51NugJ6tiRFdBoCQ-eHRt3O3Q0oZLVOoR1Smyrtw,596992
-ipex_llm/libs/libgptneox_vnni.dll,sha256=LGxqSWvTt7inWJN_2Ijc4k60Vzx2ejpr6InCpyOmbKU,568832
-ipex_llm/libs/libllama_avx.dll,sha256=sz9vGl7MLh0hMdISpeIoMAA1Dflvvv-eNYCUVu4Md50,591360
-ipex_llm/libs/libllama_vnni.dll,sha256=0PEiINTgiXh-XsHqOtF_BBkRoGvCU4RJB_ozdTGS29M,563200
-ipex_llm/libs/libstarcoder_avx.dll,sha256=HcYgXKyXLexIOT-KvUC8YbE-lmdkwzot8l5wxElQ_Ts,627712
-ipex_llm/libs/libstarcoder_vnni.dll,sha256=nsquBDWTmjuT2s96WExZFJVj8OBk7XmoaWSuo2Jo2Bc,599552
-ipex_llm/libs/llama-api.dll,sha256=1zRLPVdZYPCN-vgDtzgsxTRyZq5FGFxmFAtRa0wHAgM,25600
-ipex_llm/libs/llama.dll,sha256=5aWMlTAplsYrZ5JDgMeEQxOxKOVs1BESo2_-Yr34BAU,562688
-ipex_llm/libs/main-bloom.exe,sha256=uepQBoIYXmiBkJY-JoUXxfjUlpTHaJr_tEmKGt6uuDY,103424
-ipex_llm/libs/main-gptneox.exe,sha256=-e43xafyjHy3Lvom0R2iQJGMO_534Piwgk4JGc_Cwyc,98816
-ipex_llm/libs/main-llama.exe,sha256=2CcAdb2oZvnjxttyMPKxbD7BIaYCyZnEVXT_CQ-rSIY,99840
-ipex_llm/libs/main-starcoder.exe,sha256=hVgYvI9cjKHDZVRRis2gSahqZgH2df0ouXvjwvhjyA0,157696
-ipex_llm/libs/pipeline.dll,sha256=aTKY4oM15SW4BHVSQ7R5V2IIfUp-lwFq9JojN4yqw2k,73216
-ipex_llm/libs/quantize-bloom.exe,sha256=eGxLPTmBsLV8bgOyEsjfV2Qq5FYkD_cnVm5stL25Nbk,126464
-ipex_llm/libs/quantize-bloom_vnni.exe,sha256=sWHjYsFX1klsl7hXpOY-_Zp9et019BmSDRFqcoPExH4,128000
-ipex_llm/libs/quantize-gptneox.exe,sha256=ON4OAej6yw-yo-0-jo9Ou3pacMNl73tLGgCErlxd6vI,104448
-ipex_llm/libs/quantize-gptneox_vnni.exe,sha256=EVwFoSH4Ho8Qh4TFkB4BShLjraqNAfqUS8G1eg1btDw,104960
-ipex_llm/libs/quantize-llama.exe,sha256=qmykmdCv3PgKAzYRN4-XcPIGHlHSKwReZXjAe84DN5w,110080
-ipex_llm/libs/quantize-llama_vnni.exe,sha256=2TdT3uxYNJGm_iph-5ndPILimscAPJfOCgkEdUjfGBk,110592
-ipex_llm/libs/quantize-starcoder.exe,sha256=GXYL7iqaLwjFyUnTq1DUhOSXzlqwaZ3E2HM-oTOr5UI,127488
-ipex_llm/libs/quantize-starcoder_vnni.exe,sha256=4o9-II0U_l9C3IIRba6o1t6TO6HaLkqnShK2gSl5t5o,128512
-ipex_llm/libs/starcoder-api.dll,sha256=QICgPD98TtHymMVHngJKfA8iFAARrzGSz9X3bzbsVCs,21504
-ipex_llm/libs/starcoder.dll,sha256=ZSu4ScNOcg8LUxEA7wCOjOhxDHV5vzSehMi5FVRStXk,599040
+ipex_llm/libs/bloom-api.dll,sha256=QYogOsFgyHxoqyzV5n6AnUb8hPiQS7uQ1qdNcMnvrgY,36352
+ipex_llm/libs/bloom.dll,sha256=hLHH8IHxtpTlWYnnJCxMqr2gBqMcXv8md01UvxdJihc,507904
+ipex_llm/libs/gptneox-api.dll,sha256=0SA4owzkSEE8lgzgOYM9GWoiE3m2QCJHHxzOv4PvQrU,24576
+ipex_llm/libs/gptneox.dll,sha256=Ujy8iGpM2I19ojtcyHfm9ALrgOqgZi5v9r_fnFh5GOw,568320
+ipex_llm/libs/libbloom_avx.dll,sha256=h71kfgRtatiSVsYtqOSUY3Fq0ZAYnxgUkhQSQBSGurM,536576
+ipex_llm/libs/libbloom_vnni.dll,sha256=u6THlrq9k8BBWgc3XyBAJNmtkHusk4iHIwB6STIfUV0,508416
+ipex_llm/libs/libgptneox_avx.dll,sha256=yiOrE1HSOPoc5ILlpJPp_TfHqNdMnz56WDY60T6dcAU,596992
+ipex_llm/libs/libgptneox_vnni.dll,sha256=lgiZD7nCFdbSFDAIn7dRiCmFbYjkQ37RjS8pDvvC2MU,568832
+ipex_llm/libs/libllama_avx.dll,sha256=_1XWkqvUPjPIoyh1GWd5ylCtyBj8nlCezrePNnidbR8,591360
+ipex_llm/libs/libllama_vnni.dll,sha256=U-YtXEAlZp0atLpZ342PV3IWML2jy2u8vzXH5lL4d7E,563200
+ipex_llm/libs/libstarcoder_avx.dll,sha256=TMAf3SSQStV-44mgHqj8XCujRsDENJDgxh8TYZHRxps,627712
+ipex_llm/libs/libstarcoder_vnni.dll,sha256=LUSNgDZOz_yf83iiLfRAtiTwRcWORQX1Lco37onmDG0,599552
+ipex_llm/libs/llama-api.dll,sha256=B3x6K9CeDz0Nu_1d-rJYGOl0KajqwjIO683W_05TCBM,25600
+ipex_llm/libs/llama.dll,sha256=eaMFndnUVncngu6AG3I_PTiZG3VnNYynSLkmhQc6fEM,562688
+ipex_llm/libs/main-bloom.exe,sha256=my585BFC6Z6tdP5zh5Jg5VotNL9jcuNNa13bW3fXM9M,103424
+ipex_llm/libs/main-gptneox.exe,sha256=WVaiQtAIDSwwnysRC41_Bokrwqcze9i_paiUfeexbcA,98816
+ipex_llm/libs/main-llama.exe,sha256=p16AQ0jbFlMWSM_0HKEA2S_mHqTD_XGZ5AFpk4KZHMc,99840
+ipex_llm/libs/main-starcoder.exe,sha256=U5moV-M_5zY-h7gsC80dqBGUe3plJBnIE_E4MR2lwSU,157696
+ipex_llm/libs/pipeline.dll,sha256=dzXk9tGBOc0rfpC1WYg91AJzEMivYHCXbUdKjtTHxNE,73216
+ipex_llm/libs/quantize-bloom.exe,sha256=hpMfatW3AanzwWT1MrNH5C06XlyH2TyY7da0LheUS04,126464
+ipex_llm/libs/quantize-bloom_vnni.exe,sha256=7bffUBX-vu8UWPRPM3UqsWENyu_blqy6b-Pu4ZceuOM,128000
+ipex_llm/libs/quantize-gptneox.exe,sha256=CeOuz0r8ok1DZGp9PPh5w9-jGkn5_U7PvdsKGNEaWqY,104448
+ipex_llm/libs/quantize-gptneox_vnni.exe,sha256=I19Qj_2t9_7zx4tEXdipag-ZzZnz01Hp8iGUm9LWHgM,104960
+ipex_llm/libs/quantize-llama.exe,sha256=4H9rWcBTrEpTMQRPXZpG2JcaSrOO8vXn4z7oHGo89E0,110080
+ipex_llm/libs/quantize-llama_vnni.exe,sha256=OHb8VOiJa_oxT-pe160IfjRnJl_QPIdl2kqmBvbK2uI,110592
+ipex_llm/libs/quantize-starcoder.exe,sha256=-LOp-c3wqvEkFgCvUC9kW_EtyTjUvVayvJve56Vy8Vs,127488
+ipex_llm/libs/quantize-starcoder_vnni.exe,sha256=DoSqIE8Aitj3gS3-3zCJRnTZzyqNLRAMUSZXJv605l8,128512
+ipex_llm/libs/starcoder-api.dll,sha256=01LYwQpRMkM9xcTd4pWeAN_5lVl5isHkFeupC_tR5Ck,21504
+ipex_llm/libs/starcoder.dll,sha256=uor_SqyR_RZ7XxzSHfT4SBclBOV11TwlVX-opx0PeAY,599040
 ipex_llm/llamaindex/__init__.py,sha256=T-EbRT6GJ_8RCu-iLmSzcftOimXSPQf2d5X72AUAy2Y,874
 ipex_llm/llamaindex/llms/__init__.py,sha256=KP1lEdGqDuxPoxL1ZSH25Pm2kKMPJBWUTLR0ckSLMIU,1139
 ipex_llm/llamaindex/llms/bigdlllm.py,sha256=FQBzq1KOjfc6uofTXAha3O7TqpJkNfOFepXQmOVlbnI,26314
@@ -95,7 +95,7 @@ ipex_llm/transformers/lisa.py,sha256=F5WxbtXQ7RdKulj83h_2DnEIgKiKGZf7zvOmg6QBl2s
 ipex_llm/transformers/loader.py,sha256=c9qfJSC6-in-mkd-iKb1igk3nHWUYS3QtyH2cOazmKc,6825
 ipex_llm/transformers/lookup.py,sha256=b6OlZ9OV10R9qeWw8mVryVpDxszkjwLkldvi7GPMJY8,19614
 ipex_llm/transformers/low_bit_linear.py,sha256=1S8H684odAx5ZVDx_qNMv4FFrkJekZduao8datZYiqw,39201
-ipex_llm/transformers/model.py,sha256=FyHrEQhkHxG3FbGkhTjVOP2rgFMjc3AXcjDwvvB0HqU,40798
+ipex_llm/transformers/model.py,sha256=r9jLsLlGB4rOZ7SMrmCE5qyXqC9Ae56unNq1gNjfP54,40885
 ipex_llm/transformers/modelling_bigdl.py,sha256=7JpNVMuyq_OmtNUaMFMXdxPWZp2q0QHC02QeA-VTPOw,6709
 ipex_llm/transformers/npu_model.py,sha256=zgXOiLIJ-3p-1Kejgv4jUFK8OiBZbezMZrRyn0_6_8c,40306
 ipex_llm/transformers/patches.py,sha256=G9KcXxo42H1HJEDaroq4JbBN5P0P0lty7U7kk7-g4tw,991
@@ -148,7 +148,7 @@ ipex_llm/transformers/models/chatglm4.py,sha256=QvUehdaCePB3MNHyWg3dneDxmjtBdxYe
 ipex_llm/transformers/models/chatglm4v.py,sha256=Ba9Xtzwtzk_rzg5khGqDrlHfJsDwc5YcM5_yPoord7o,13324
 ipex_llm/transformers/models/common.py,sha256=ueLGko8May2qWdjI-lSH30LXY4NYrqtBDXZekfq9rfQ,15374
 ipex_llm/transformers/models/decilm.py,sha256=P-PBuDPf07GvKggLwJx_wPwIn6esN3rX8ai2JxRuZmE,5246
-ipex_llm/transformers/models/deepseek.py,sha256=LFFA3tOtS3-WVoNmQQsfAymMCwNZiVQyBRZ4ZN_-IaE,13135
+ipex_llm/transformers/models/deepseek.py,sha256=BJocxhznzvM99IZeAWuhyHEBm6Z2-AwhLFYMdjMJuc4,13083
 ipex_llm/transformers/models/deepseek_v3.py,sha256=CTgwIKQlUPlUCbOxc9Id5GapWkXOP6pMtkguYrWpCio,10003
 ipex_llm/transformers/models/gemma.py,sha256=_E3Yw8Y45xyNVeLqyVKcpr8kjuICtETeL82cJ-bWJuU,9424
 ipex_llm/transformers/models/gemma2.py,sha256=2WZuv-FLzJyTJFaYxOuzJt47QE64M0lHnzAiO5T6ozI,8049
@@ -180,7 +180,7 @@ ipex_llm/transformers/models/rwkv5.py,sha256=OkRNj1pCAZg1z2Fw-I0DEnxLEdZyPeRSQ6m
 ipex_llm/transformers/models/sd.py,sha256=VvHV5u-0k2MgHu3NL9113hPj7DgfxqctuKzEEeNfRDU,5981
 ipex_llm/transformers/models/stablelm.py,sha256=fj-XtOnR6kggnFUQTMPCOOzolkPztN06WAv8QW-XRnI,7054
 ipex_llm/transformers/models/starcoder2.py,sha256=ONKvD7JCkRM0DI-R56x28QFBJ7CjD5hOZBQ_3WfOcNk,6626
-ipex_llm/transformers/models/utils.py,sha256=Rj7QK1s3QxsTT4HzGNR00Q7izG5xYiZEX29YoQIM1tA,16998
+ipex_llm/transformers/models/utils.py,sha256=ch1dpIegGL-jxIAS9OdkojS2qNXRwhebHujVwGH5Y0o,17266
 ipex_llm/transformers/models/whisper.py,sha256=ju3WP8Eq-KvD7kb3Qy51r4FOfSX3NBxfp5RBcq__gzc,4241
 ipex_llm/transformers/models/yuan.py,sha256=JYAn_ZaSGK0NBJLEIxCACfAq084a66GFJkdd5NbpmMA,7732
 ipex_llm/transformers/npu_models/__init__.py,sha256=ulEUGLjaP48LCrVeury3UxLjXxKzRi0UpSG4bYu-7f8,585
@@ -256,11 +256,11 @@ ipex_llm/vllm/xpu/engine/__init__.py,sha256=pY_CpyuZd72fr6s32ejeKHKFW0K4vUU2rzZj
 ipex_llm/vllm/xpu/engine/engine.py,sha256=NvCMbp0X8NVrOqbwm4FTvXOptTRLzu9jQsy37ZHnTk8,9493
 ipex_llm/vllm/xpu/entrypoints/openai/api_server.py,sha256=IjiSze9vzBCAkLu_VwIcJwuO1jyFna7DLrj6aSL7RaQ,35220
 ipex_llm/vllm/xpu/entrypoints/openai/cli_args.py,sha256=hB398yYtKauASRzevctScdbFIjiiSGMAe1bwEuIHrhY,10893
-ipex_llm-2.3.0b20250421.data/scripts/ipex-llm-init.bat,sha256=HPtCYuDYwEatq7dAwOvdfVcHYCpAVdbj75K1qh0vQek,2578
-ipex_llm-2.3.0b20250421.data/scripts/llm-chat.ps1,sha256=6qrs-hGVAV8IKh7Jx8nq_XrnZcjd7qGU5wndArM7Yag,2769
-ipex_llm-2.3.0b20250421.data/scripts/llm-cli.ps1,sha256=3qBtTLs_EjYDnM8YyCpJhzLnGCKTEGssu9UNqfkjVXs,3009
-ipex_llm-2.3.0b20250421.dist-info/METADATA,sha256=ipsTM8bckoMMOBFjsTPqRqjth7Z5JAKQBSaOoMRlL4g,13917
-ipex_llm-2.3.0b20250421.dist-info/WHEEL,sha256=6iYPr8vTHsyDK75jr9X0V3I9wPSVmtwr_8fdATBciGk,98
-ipex_llm-2.3.0b20250421.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
-ipex_llm-2.3.0b20250421.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
-ipex_llm-2.3.0b20250421.dist-info/RECORD,,
+ipex_llm-2.3.0b20250422.data/scripts/ipex-llm-init.bat,sha256=HPtCYuDYwEatq7dAwOvdfVcHYCpAVdbj75K1qh0vQek,2578
+ipex_llm-2.3.0b20250422.data/scripts/llm-chat.ps1,sha256=6qrs-hGVAV8IKh7Jx8nq_XrnZcjd7qGU5wndArM7Yag,2769
+ipex_llm-2.3.0b20250422.data/scripts/llm-cli.ps1,sha256=3qBtTLs_EjYDnM8YyCpJhzLnGCKTEGssu9UNqfkjVXs,3009
+ipex_llm-2.3.0b20250422.dist-info/METADATA,sha256=Gs6zvDGgubYwrRm9k94z1kKksr8QoqJu4Gv58C_gKj4,8865
+ipex_llm-2.3.0b20250422.dist-info/WHEEL,sha256=6iYPr8vTHsyDK75jr9X0V3I9wPSVmtwr_8fdATBciGk,98
+ipex_llm-2.3.0b20250422.dist-info/entry_points.txt,sha256=TiUyBB2MRmfF3ko-pyAEzqeBCRnyhu27bNOAsWPp3e8,61
+ipex_llm-2.3.0b20250422.dist-info/top_level.txt,sha256=CGCMHM-SyqUabU4h8RqJ2KTYckQUO3LvIWwmUQ6Qbzw,9
+ipex_llm-2.3.0b20250422.dist-info/RECORD,,