jpSSKrSSKrSSKrSSKJr SSKJrJrJr SSK J r J r J r SSK JrJr \R "\5r"SS\\5r"SS \5rg) N)Enum) AnnotatedAnyUnion)Fieldfield_validatormodel_validator) BaseSettingsSettingsConfigDictc,\rSrSrSrSrSrSrSrSr Sr g ) AcceleratorDevicezDevices to run model inferenceautocpucudampsxpuN) __name__ __module__ __qualname____firstlineno____doc__AUTOCPUCUDAMPSXPU__static_attributes__rb/home/mande/repo/quber/.venv/lib/python3.13/site-packages/docling/datamodel/accelerator_options.pyr r s( D C D C Cr r c\rSrSr%Sr\"SSSS9rSr\\ \ "SS 94\ S 'S r \\ \\4\ "S S 94\ S 'Sr\\\ "SS 94\ S'\"S 5S5r\"SS9\S\S\4Sj55rSrg)AcceleratorOptionsz|Hardware acceleration configuration for model inference. Can be configured via environment variables with DOCLING_ prefix. DOCLING__T) env_prefixenv_nested_delimiterpopulate_by_nameaNumber of CPU threads to use for model inference. Higher values can improve throughput on multi-core systems but may increase memory usage. Can be set via DOCLING_NUM_THREADS or OMP_NUM_THREADS environment variables. Recommended: number of physical CPU cores.) description num_threadsraHardware device for model inference. Options: `auto` (automatic detection), `cpu` (CPU only), `cuda` (NVIDIA GPU), `cuda:N` (specific GPU), `mps` (Apple Silicon), `xpu` (Intel GPU). Auto mode selects the best available device. Can be set via DOCLING_DEVICE environment variable.deviceFaEnable Flash Attention 2 optimization for CUDA devices. Provides significant speedup and memory reduction for transformer models on compatible NVIDIA GPUs (Ampere or newer). Requires flash-attn package installation. Can be set via DOCLING_CUDA_USE_FLASH_ATTENTION2 environment variable.cuda_use_flash_attention2cU[Vs1sHo"RiM sn;d[R"SU5(aU$[ S5es snf)Nz ^cuda(:\d+)?$zLInvalid device option. Use `auto`, `cpu`, `mps`, `xpu`, `cuda`, or `cuda:N`.)r valuerematch ValueError)clsr0ds r!validate_device"AcceleratorOptions.validate_deviceHsS &78&7WW&78 8BHH e= = L   9sAbefore)modedatareturnc,[U[5(aWURS5nUcC[R"S5n[R"S5nUcUb[ U5US'U$U$![ a [RSU5 U$f=f)a Set num_threads from the "alternative" envvar OMP_NUM_THREADS. The alternative envvar is used only if it is valid and the regular envvar is not set. Notice: The standard pydantic settings mechanism with parameter "aliases" does not provide the same functionality. In case the alias envvar is set and the user tries to override the parameter in settings initialization, Pydantic treats the parameter provided in __init__() as an extra input instead of simply overwriting the evvar value for that parameter. r,DOCLING_NUM_THREADSOMP_NUM_THREADSz1Ignoring misformatted envvar OMP_NUM_THREADS '%s') isinstancedictgetosgetenvintr3_logerror)r4r:input_num_threadsdocling_num_threadsomp_num_threadss r!check_alternative_envvars,AcceleratorOptions.check_alternative_envvarsTs dD ! ! $ 7  (&(ii0E&F#"$)),=">&.?3N.1/.B]+  t & O+ sA// BBrN)rrrrrr model_configr,rrDr__annotations__r-rstrr r.boolrr6r classmethodrrJrrr r!r#r#s &C$L   &     .  I c$$% ;    . y  J    X    (#SS$r r#)loggingrBr1enumrtypingrrrpydanticrrr pydantic_settingsr r getLoggerrrErNr r#rr r!rWsN ((<<>"TWWr