Ki{CJSrSSKrSrSrSSKJrJrJrJrJ r J r J r J r J r JrJrJrJrJrJrJrJrJrJrJrJrJrJrJrJrJrJrJ r J!r!J"r"J#r#J$r$J%r%J&r&J'r'J(r(J)r)J*r*J+r+J,r,J-r-J.r. Sr/SSK2J3r3 \/(a\/eSSK4J5r5J6r6J7r7J8r8J9r9J:r:J;r;Jr> \3R"5urArBrC\B(a\Br\3R"5 S \E4S jrFSS \ES \ES-4S jjrGS\ES\E4SjrHS\ES\E4SjrISS\JS\JS\J4SjjrKSrLSS\JS\JS\J4SjjrMg!\0a r1\1r/Sr1C1NSr1C1ff=f!\?a Nf=f)a ONNX Runtime is a performance-focused scoring engine for Open Neural Network Exchange (ONNX) models. For more information on ONNX Runtime, please see `aka.ms/onnxruntime `_ or the `Github project `_. Nz1.24.2 Microsoft)* ExecutionModeExecutionOrderGraphOptimizationLevel LoraAdapter ModelMetadataNodeArgOrtAllocatorType OrtArenaCfgOrtCompileApiFlagsOrtDeviceMemoryTypeOrtEpAssignedNodeOrtEpAssignedSubgraph OrtEpDevice OrtExecutionProviderDevicePolicyOrtExternalInitializerInfoOrtHardwareDeviceOrtHardwareDeviceType OrtMemoryInfoOrtMemoryInfoDeviceType OrtMemTypeOrtSparseFormat OrtSyncStream RunOptionsSessionIOBindingSessionOptionscreate_and_register_allocator create_and_register_allocator_v2disable_telemetry_eventsenable_telemetry_eventsget_all_providersget_available_providersget_build_info get_deviceget_ep_devicesget_version_stringhas_collective_ops#register_execution_provider_libraryset_default_logger_severityset_default_logger_verbosityset_global_thread_pool_sizesset_seed%unregister_execution_provider_library)onnxruntime_validation) AdapterFormatInferenceSession IOBinding ModelCompiler OrtDeviceOrtValue SparseTensor copy_tensors) experimental package_namecFSSKJnJn U"U5nU$!Ua SnU$f=f)Nr)PackageNotFoundErrorversion)importlib.metadatar;r<)r9r;r<package_versions V/var/www/html/dynamic-report/venv/lib/python3.13/site-packages/onnxruntime/__init__.py_get_package_versionr@js8@!,/   s   directory_namecSSKJnJn U=(d UnU"U5nUR=(d /nUHPnURR S5(dM%XGR ;dM6UR5Rs $ U(dEUH?nURR S5(dM%UR5Rs $ g!Ua gf=f)Nr)r; distributionz __init__.py) r=r;rCfilesnameendswithpartslocateparent)r9rAr;rCroot_directory_namedistrDfiles r?_get_package_rootrMtsE(8L L)  bDyy!!-005HJJ5V{{}+++ 99%%m44;;=///      s0AC C*C,C6CCCC version_strreturnc<U(aURS5S$S$)zExtract CUDA major version from version string (e.g., '12.1' -> '12'). Args: version_str: CUDA version string to parse Returns: Major version as string, or "12" if parsing fails .r12)split)rNs r?_extract_cuda_major_versionrTs!)4;  S !! $== cuda_majorcUS:XaS$S$)zGet cufft library version based on CUDA major version. Args: cuda_major: CUDA major version as string (e.g., "12", "13") Returns: cufft version as string 13rR11)rVs r?_get_cufft_versionr[s%4/4/rU is_windowscudacudnnc ([[5n[U5nU(a+SSSSUS34SSSSUS34SSSSUS34SS SS US34/n/S QnO.SSS S U34SSS SU34SSS SU34SSSS SU34SS S SU34/nS/nU(aUO/U(aU-$/-$)Nnvidiacublasbin cublasLt64_z.dll cublas64_cufftcufft64_ cuda_runtime cudart64_))r`r^rbz&cudnn_engines_runtime_compiled64_9.dll)r`r^rbz!cudnn_engines_precompiled64_9.dll)r`r^rbzcudnn_heuristic64_9.dll)r`r^rbzcudnn_ops64_9.dll)r`r^rbzcudnn_adv64_9.dll)r`r^rbzcudnn_graph64_9.dll)r`r^rbz cudnn64_9.dlllibzlibcublasLt.so.z libcublas.so. cuda_nvrtcz libnvrtc.so.)r`curandrizlibcurand.so.10z libcufft.so.z libcudart.so.)r`r^riz libcudnn.so.9)rT cuda_versionr[)r\r]r^cuda_major_version cufft_versioncuda_dll_pathscudnn_dll_pathss r?_get_nvidia_dll_pathsrqs4\B&'9:Mx+6H5I(N O x)4F3Gt(L M w-'E F ~u :L9MT.R S    x/:L9M(N O x-8J7K(L M |Ul;M:N,O P : wm_'E F ~u >P=Q.R S   8 #N% PPR PPrUc SSKnSSKnSSKnSSKJn [ [ S[35 [(a[ S[35 [ SUR55 [ S5 /nU"5HznURSnUS :XdURS 5(dM0[US 5nU(dMEXd;dMLURU5 [ US URS U35 M| [U5S :a [ S5 [(a[[5nSSU3SU3SU3SU3SU3SU3SU3/n U H[nURS5(aSOSn [!U5n U (a[ US U S [Xj535 MM[ US35 M] UR#5S:Xa)[ SUR$R'SS535 O# UH oT;v M g7fNrZ).0keywordr|s r? :print_debug_info..is_target_dll.. sFo7$os)rlany)r|target_keywordss` r? is_target_dll'print_debug_info..is_target_dlls<-z:O|#*HhSZ"m]l"mFoFF FrUz List of loaded DLLs:cpuinfopy3nvml)get_device_infoz Device information:zCplease `pip install py-cpuinfo py3nvml` to show device information.z0please `pip install psutil` to show loaded DLLs.)importlib.utilosplatformr=rsprintr9 __version__rlmetadata startswithrMappendr<lenrTr@systemenvirongetutil find_specstrr{Processgetpid memory_mapsr|lowertransformers.machine_infor) importlibrrrs ort_packagesrKpackagelocationrVpackagesrAr<rr{prirs r?print_debug_infors 0 \N*[M 23| ,\N;< +x((*+ 34L--' m #w'9'9:R'S'S(-@HxW8##G, DLL>hZ@A  <1  e |0>  $ZL 1j\ *zl +j\ *zl +":, /!* .   G)0););I)F)FXDN*73G G9D1B71[0\]^ 01  I% .rzz~~fi/P.QRS  g % 9"**..IZ\e:f9ghi~~)) G G  NN299; ' &'==?CSXX^^-..chh# <~~'' 22y~~7O7OPY7Z7ZF-.o'([\  @ArUmsvcc SSKnSSKnSSKnSSKnUR 5S;agUR 5S:HnU(aOU(aHUR S5 UR S5 UR 5S:waUR S5 Sn [(aR[[RS 5S5n U S :a)U(dU(a[S [S[S35 gOU(dU(agSn U(a[S5n Sn U (aKSU ;aE[R "[5 U RS5Sn [U SS5n SSS5 U (a U (aX:HOSnSUR";a4Un U (a+U (a$X:wa[SU SU S[SU SU S3 5 U(a2Uc/[%SS5nU(aUR&R)US5nU=(d SnUR&R+U5(d9UR&R)UR&R-[.5U5nUR&R1U5nUR&R3U5(d[5SUS35eU (a [S 5 g[7XU5n/nUHnU(aUR&R)UUS!5OUR&R("U/UQ76nUR&R9U5(dMhUR U5nUR;US!5 M SnUH"nUS!nUU;dMUR U5nM$ U(a [S%5 gg![a [S 5 [S 5 GN+f=f![a [S[S35 gf=f!,(df  GN=f![<an[S"US#U35 SnAGMNSnAff=f![<anS$n[S"US#U35 SnAMSnAff=f)&aPreload CUDA 12.x+ and cuDNN 9.x DLLs in Windows or Linux, and MSVC runtime DLLs in Windows. When the installed PyTorch is compatible (using same major version of CUDA and cuDNN), there is no need to call this function if `import torch` is done before `import onnxruntime`. Args: cuda (bool, optional): enable loading CUDA DLLs. Defaults to True. cudnn (bool, optional): enable loading cuDNN DLLs. Defaults to True. msvc (bool, optional): enable loading MSVC DLLs in Windows. Defaults to True. directory(str, optional): a directory contains CUDA or cuDNN DLLs. It can be an absolute path, or a path relative to the directory of this file. If directory is None (default value), the search order: the lib directory of compatible PyTorch in Windows, nvidia site packages, default DLL loading paths. If directory is empty string (""), the search order: nvidia site packages, default DLL loading paths. If directory is a path, the search order: the directory, default DLL loading paths. rN)rwryrwzvcruntime140.dllz msvcp140.dllARM64zvcruntime140_1.dllz]Microsoft Visual C++ Redistributable is not installed, this may lead to the DLL load failure.zGIt can be downloaded at https://aka.ms/vs/17/release/vc_redist.x64.exe.rQ zWARNING: z is built with CUDA z, which is not supported for preloading. CUDA 12.x or newer is required. Call preload_dlls with cuda=False and cudnn=False.z,WARNING: Unable to parse CUDA version 'zR'. Skipping DLL preloading. Call preload_dlls with cuda=False and cudnn=False.Frvz+cur7z$WARNING: The installed PyTorch z uses CUDA z.x, but z$.x. Please install PyTorch for CUDA z.x to be compatible.riz..zInvalid parameter of directory=z. The directory does not exist!z9Skip loading CUDA and cuDNN DLLs since torch is imported.zFailed to load z: Tz\Please follow https://onnxruntime.ai/docs/install/#cuda-and-cudnn to install CUDA and CuDNN.)ctypesrrsysrCDLLmachineOSErrorrrlintrSr9 ValueErrorr@ contextlibsuppressmodulesrMr|joinisabsdirname__file__normpathisdir RuntimeErrorrqisfiler Exception)r]r^r directoryrrrrr\ort_cuda_majoris_cuda_cudnn_imported_by_torch torch_versiontorch_cuda_majorcu_partis_torch_cuda_compatible torch_rootbase_directory dll_paths loaded_dlls relative_pathdll_path_e has_failure dll_filenames r? preload_dllsr#s"  44"i/Jd ] KK* + KK '!W, 01 N|  !3!3C!8!;IK"  GGLLr): ;n=}=  77>>( # # 9KK)""=#45#K" $R( { * =KK - # lm} ] q r [ \ ]  A,Pee    "10` 9zA3788 9 ="  ~Rs;<< =sbANAN>:#O!0%O3.P N;:N;>OO! O03 P=PP Q'QQr)TT)TTTN)N__doc__rr __author__onnxruntime.capi._pybind_staterrrrrr r r r r rrrrrrrrrrrrrrrrrrr r!r"r#r$r%r&r'r(r)r*r+r,r-import_capi_exceptionrronnxruntime.capir.1onnxruntime.capi.onnxruntime_inference_collectionr/r0r1r2r3r4r5r6r8 ImportError!get_package_name_and_version_infor9r<rlcheck_distro_inforr@rMrTr[boolrqrrrZrUr?rsl    0+++++++++++Z!4      '=&^&^&`# g| K((*sCt0 >S >S > 03 03 0&Qd&Q$&Qd&QRRBjntn4ndnw,  s*AD DD DDD"!D"