jRbSrSSKrSSKrSSKJr SSKJr SSKJr SSK J r J r J r J r Jr SSKrSSKrSSKrSSKJrJr SS KJrJr SS KJrJr SS KJr SS KJr SS K J!r!J"r" SSK#J$r$ \RJ"\&5r'Sr(Sr)"SS5r*SVSjr+SSSSSSSSSSSSSSSSSSS\\SSSSSSSS\RX\RZ"S5SSSSSSSSSSSSSS4-S \\\4S!\\.\ \.\.4\ \.\.\.44S"\.S#\/S$\/S%\0S&\1S'\.S(\/S)\ \1S*\ \ \0\04S+\ \ \0\04S,\0S-\0S.\0S/\ \0S0\0S1\0S2\ \1S3\.S4\.S5\1S6\ \0S74S8\ \0S74S9\.S:\/S;\ \0S<\ \1S=\ \.S>\ \ S?\/S@\/SA\RdSB\RZSC\/SD\/SE\/SF\1SG\/SH\ \ \\.\ \.\.4\ \.\.\.44SI\ \ \.SJ\ \ \0SK\.SL\1SM\ \.SN\0SO\0SP\ \.4`SQjjr3"SRSS\RhRjRl5r7"STSU\85r9g)WaLoader Factory, Fast Collate, CUDA Prefetcher Prefetcher and Fast Collate inspired by NVIDIA APEX example at https://github.com/NVIDIA/apex/commit/d5e2bb4bdeedd27b1dfaf5bb2b24d6c000dee9be#diff-cf86c282ff7fba81fad27a559379d5bf Hacked together by / Copyright 2019, Ross Wightman N)suppress)partial)repeat)CallableOptionalSequenceTupleUnion)IMAGENET_DEFAULT_MEANIMAGENET_DEFAULT_STD)IterableImageDataset ImageDataset)OrderedDistributedSamplerRepeatAugSampler) RandomErasing)FastCollateMixup)ScheduledBatchSamplerScheduledTransformDataset)create_transformc N[US[5(de[U5n[USS[5(Ga,[USS[R5n[USS5nX-n[ R "U[ RS9n[ R "U/USSSRQ7[ RS9n[U5Hn[XS5U:Xde[U5H`nXSXWX--'U(a/XgX--==[ R"XSU5- ss'MGXgX--==XSU- ss'Mb M Xe4$[USS[R5(a[ R"UV s/sHoSPM sn [ RS9n[U5U:Xde[ R "U/USSRQ7[ RS9n[U5H(nXg==[ R"XS5- ss'M* Xe4$[USS[ R5(a[ R"UV s/sHoSPM sn [ RS9n[U5U:Xde[ R "U/USSRQ7[ RS9n[U5HnXgRXS5 M Xe4$es sn fs sn f)zcA fast collation function optimized for uint8 images (np array or torch) and int64 targets (labels)r)dtyper ) isinstancetuplelennpndarraytorchzerosint64shapeuint8range from_numpytensorTensorcopy_) batch batch_sizeis_npinner_tuple_sizeflattened_batch_sizetargetsr%ijbs ڞC:\Users\johnn\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.13_qbz5n2kfra8p0\LocalCache\local-packages\Python313\site-packages\timm/data/loader.py fast_collater2s eAh & && &UJ%(1+u%%58A; 3uQx{+)<++2%++F2JU1Xa[^5I5IJRWR]R]^z"Aux{#'77 77+,.3hqkAN*+q~-.%2B2B58A;q>2RR.q~-.%(1+a.@. -# E!HQK , ,,,e4e!e4EKKH7|z)))j=58A;+<+<=U[[Qz"A I))%(1+6 6I# E!HQK . .,,e4e!e4EKKH7|z)))j=58A;+<+<=U[[Qz"A IOOEHQK (#u 5 5s 'LL"c<[U[[45(d[[X55nU$[ U5U:waE[ R "U5R5nU4U-n[RSUS35 U$[ U5U:XdS5eU$)Nz@Pretrained mean/std different shape than model, using avg value .z-normalization stats must match image channels) rrlistrrrmeanitem_loggerwarning)xnx_means r1 adapt_to_chsr=Es a% ' ' &,  H Q1" IMZ[\Z]]^_` H1v{KKK{ Hcp\rSrSr\\S\R"S5SSSSSS 4 S \RRRS \ \ S 4S \ \ S 4S\ S\RS\\RS\S\ S\S\ S\ 4SjjrSrSr\S5r\S5r\S5r\S5r\R4S5rSrg)PrefetchLoaderQcudaNFconstr rloaderr6.stdchannelsdevice img_dtypefp16re_probre_modere_count re_num_splitsc [X$5n[X45nSUSS4n XlXPlU(a[RnU=(d [R Ul[R"UV s/sHoS-PM sn XVS9RU 5Ul [R"UV s/sHoS-PM sn XVS9RU 5Ul US:a[UU U U US9Ul OSUl URS:H=(a [RR5UlURS:H=(a [R"R5Ulgs sn fs sn f)Nr )rIrrD) probabilitymode max_count num_splitsrIrCnpu)r=rFrIrfloat16float32rJr%viewr6rGrrandom_erasingtyperC is_availableis_cudarVis_npu)selfrFr6rGrHrIrJrKrLrMrNrOnormalization_shaper:s r1__init__PrefetchLoader.__init__Ss0D+3) (Aq1   I"3emmLL" #dWd #FEEITJ]E^ <<! "cWc "6DDHDI\D]  R<"/#"( #D #'D {{f,J1H1H1J kkU*Guyy/E/E/G  $ "s 2E).E.c## SnUR(aJ[RRURS9n[ [RR US9nOcUR(aJ[RRURS9n[ [RR US9nOSn[nURGH[upEU"5 URURSS9nURURSS9nURUR5RUR5RUR 5nUR"bUR#U5nSSS5 U(dWW4v OSnUbUR(a7[RR%URS9R'U5 OGUR(a6[RR%URS9R'U5 UnUnGM^ WW4v g!,(df  N=f7f)NT)rI)stream)rI non_blockingF)r]rrCStreamrIrrdr^rVrrFtorJsub_r6div_rGrZcurrent_stream wait_stream)r_firstrdstream_context next_input next_targetinputtargets r1__iter__PrefetchLoader.__iter__|s <<ZZ&&dkk&:F$UZZ%6%6vFN [[YY%%T[[%9F$UYY%5%5fENF%N'+{{ #J!']]$++D]Q )nnDKKdnS ']]4>>:?? JOOPTPXPXY &&2!%!4!4Z!@J "Vm#!<<JJ--T[[-AMMfU[[II,,DKK,@LLVTE F+(3.Vm+"!s CIB H6c.URR$ru)rFsamplerrvs r1rzPrefetchLoader.sampler{{"""r>c.URR$ru)rF batch_samplerrvs r1r~PrefetchLoader.batch_samplers{{(((r>c.URR$ru)rFdatasetrvs r1rPrefetchLoader.datasetr|r>c[URR[5(a URRR$g)NFrrF collate_fnr mixup_enabledrvs r1rPrefetchLoader.mixup_enableds3 dkk,,.> ? ?;;))77 7r>c[URR[5(aXRRlggrur)r_r:s r1rrs0 dkk,,.> ? ?34KK " " 0 @r>)rIrJr]r^rFr6rZrG)__name__ __module__ __qualname____firstlineno__r r rrIutilsdata DataLoaderr floatintrrboolstrrarrrwpropertyrzr~rrsetter__static_attributes__r>r1r@r@QsM '<%9#(<<#7/3"!"'HKK$$//'Hs #'Hucz" 'H  'H LL 'H  ,'H'H'H'H'H'HR#J ##))## 55r>r@allc[RRR5nURU:Xde[ U[ 5(aWU"U5n[R"U5 [R"U5 [RRUS-5 gUS;deUS:Xa-[RRURS-5 gg)Nl)rpartr) rrrget_worker_infoidrrrandomseed manual_seedr) worker_idworker_seeding worker_infors r1 _worker_initrs++""224K >>Y && &.(++k* D $ t{+,000 U " IINN;++{; < #r>FrDrEg?g?bilinearrCTconstantg?g?r input_sizer) is_trainingno_augrLrMrNre_splittrain_crop_modescaleratiohflipvflip color_jittercolor_jitter_probgrayscale_probgaussian_blur_prob auto_augmentnum_aug_repeatsnum_aug_splits interpolationr6.rG num_workers distributedcrop_pct crop_modecrop_border_pixelsr pin_memoryrKrJrIuse_prefetcheruse_multi_epochs_loaderpersistent_workersrtf_preprocessinginput_size_choicesbatch_size_choicesbatch_choice_weightsbatch_choice_seedbatch_choice_schedulebatch_schedule_epochsbatch_schedule_spreadbatch_schedule_random_mix num_batchesc0 Sn0U(a U=(d Sn0[S40SU_SU_SU _SU _SU _SU _S U _S U_S U_S U_S U_SU_SU_SU_SU_SU_SU_SU_SU_SU_SU_SU0_SU&_SU"_SUS:_6n1[U[[45(a[ U5S:XaUSO [ U5n2U'SLn3U3(GaU(d [ S5eU#(a [ S5e[U[ RRR5(a [S 5eUS:a [ S!5eU'(d [ S"5e/n4U'Hn5[U5[5(aU2U5U54n5OP[ U55S:XaU2/U5Q7n5O:[ U55S:Xa [U55n5U5SU2:wa [ S#5eO [ S$5e[S%U555(a [ S&5eU4RU55 M U(cU/[ U45-n(O#[ U(5[ U45:wa [ S'5eU)b#[ U)5[ U45:wa [ S(5eU4V5s/sHn5[U540U1D6PM n6n5SUl[!UU65nO[U40U1D6Ul[U["5(aUR%US)9 Sn7U(a[U[ RRR5(dYU(aFU(a ['UUS*9n7OM[ RRR(R+U5n7O[-U5n7O US:XdS+5eU3(a,U7c)[ RRR/U5n7Uc;U"(a[0O-[ RRR2R4n[ RRR6n8U#(a[8n8[UUU[;[<U%S,9U$=(a US:S-9n9U3(a[?U7U(U)U*UU/U+U,U-U.S.9 U9S/'OWU9RAU[U[ RRR5(+=(a U7SL=(a UU7US09 U8"U40U9D6n:U"(a%U(a U(dUOS2n<[EU:UUU2U!UU U zoom in). ratio: Random aspect ratio range (crop ratio for RRC, ratio adjustment factor for RKR). hflip: Horizontal flip probability. vflip: Vertical flip probability. color_jitter: Random color jitter component factors (brightness, contrast, saturation, hue). Scalar is applied as (scalar,) * 3 (no hue). color_jitter_prob: Apply color jitter with this probability if not None (for SimlCLR-like aug grayscale_prob: Probability of converting image to grayscale (for SimCLR-like aug). gaussian_blur_prob: Probability of applying gaussian blur (for SimCLR-like aug). auto_augment: Auto augment configuration string (see auto_augment.py). num_aug_repeats: Enable special sampler to repeat same augmentation across distributed GPUs. num_aug_splits: Enable mode where augmentations can be split across the batch. interpolation: Image interpolation mode. mean: Image normalization mean. std: Image normalization standard deviation. num_workers: Num worker processes per DataLoader. distributed: Enable dataloading for distributed training. crop_pct: Inference crop percentage (output size / resize size). crop_mode: Inference crop mode. One of ['squash', 'border', 'center']. Defaults to 'center' when None. crop_border_pixels: Inference crop border of specified # pixels around edge of original image. collate_fn: Override default collate_fn. pin_memory: Pin memory for device transfer. fp16: Deprecated argument for half-precision input dtype. Use img_dtype. img_dtype: Data type for input image. device: Device to transfer inputs and targets to. use_prefetcher: Use efficient pre-fetcher to load samples onto device. use_multi_epochs_loader: persistent_workers: Enable persistent worker processes when ``num_workers`` is greater than zero. worker_seeding: Control worker random seeding at init. tf_preprocessing: Use TF 1.0 inference preprocessing for testing model ports. input_size_choices: Per-batch input size choices for scheduled resolution training. batch_size_choices: Batch size corresponding to each input size choice. Uses ``batch_size`` for all choices when not specified. batch_choice_weights: Sampling weights for input size choices. Uniform when not specified. batch_choice_seed: Random seed used to create and shuffle the batch schedule. batch_choice_schedule: Choice schedule mode, either ``constant`` or ``progressive``. batch_schedule_epochs: Number of epochs over which a progressive schedule moves through its choices. batch_schedule_spread: Standard deviation of the progressive choice window, in choice-index units. batch_schedule_random_mix: Fraction of uniform random exploration over nonzero-weight choices in a progressive schedule. num_batches: Fixed number of loader batches per epoch. Inferred from the schedule-average batch size for progressive schedules when not specified. Returns: DataLoader rrrrrrrrrrrrrrr6rGrrrrLrMrNrOrrseparaterBNz>Scheduled input sizes are only supported for training loaders.zBMultiEpochsDataLoader is not supported with scheduled input sizes.z2Scheduled input sizes require a map-style dataset.zAAugmentation splits are not supported with scheduled input sizes.z2input_size_choices must contain at least one size.z?All scheduled input sizes must use the same number of channels.z@Scheduled input sizes must be scalars, HW tuples, or CHW tuples.c3*# UH oS:*v M g7f)rNr).0 dimensions r1 create_loader..ws84i>4sz5All scheduled input size dimensions must be positive.zDbatch_size_choices and input_size_choices must have the same length.zFbatch_choice_weights and input_size_choices must have the same length.)r) num_repeatszORepeatAugment not currently supported in non-distributed or IterableDataset use)r)rrrworker_init_fnr) batch_sizeschoice_weightsr drop_lastrchoice_scheduleschedule_epochsschedule_spreadschedule_random_mixr~)r)shufflerzrrrD) r6rGrHrIrKrJrLrMrNrOr)#dictrrr5r ValueErrorrrrIterableDataset TypeErrorranyappendr transformrrset_loader_cfgrrDistributedSamplerr RandomSamplerr2 dataloaderdefault_collaterMultiEpochsDataLoaderrrrupdatepopr@)=rrr)rrrLrMrNrrrrrrrrrrrrrrr6rGrrrrrrrrKrJrIrrrrrrrrrrrrrrrOtransform_kwargsrHscheduled_batchingresolved_input_sizessize transformsrz loader_class loader_argsrFeprefetch_re_probs= r1 create_loaderrsMXM&+! (     ",&."$  !"#$.%&'()*+,$-.*/0&12 !#3: j5$- 0 0S_5I 1 Y ,47]^ ^ "ab b gu{{//?? @ @PQ Q A `a a!QR R!&D$$$ $-Ta (4(TaT{7h&$%fgg'!!cdd84888 !XYY ' ' -'  %",4H0I!I  # $,@(A Acd d  +4H0ISQeMf0fef fMabMaT&t@/?@Ma b +GZ@,ZL;KL'/00 ;7G:gu{{/?/?/O/OPP *7P++**66II'R08G!#v%vv#go++""009%3\9I9I9T9T9d9d ;;##..L, |NK-A+/ K'< */"!#111 9 ( O$ !"7EKK,<,<,L,LMMqRY]aRaqfq!  6g55&1&7b $'   Mgc@ 6,-g556sR6? R;; S)S$$S)c4^\rSrSrU4SjrSrSrSrU=r$)ric>[TU]"U0UD6 SUlURc[ UR 5UlO[ UR5UlSUl[TU]5Ulg)NFT)superra_DataLoader__initializedr~_RepeatSamplerrzrriterator)r_argskwargs __class__s r1raMultiEpochsDataLoader.__init__sc $)&)(-%    %)$,,7DL!/0B0B!CD (,%(* r>cURc[UR5$[URR5$ru)r~rrzrvs r1rwMultiEpochsDataLoader.__len__s2$($6$6$>s4<< cCHZHZHbHbDccr>c#p# [[U55Hn[UR5v M g7fru)r#rnextr)r_r.s r1rrMultiEpochsDataLoader.__iter__s's4y!At}}% %"s46)rr~rrz) rrrrrarwrrr __classcell__)rs@r1rrs+d&&r>rc$\rSrSrSrSrSrSrg)riz;Sampler that repeats forever. Args: sampler (Sampler) cXlgrurz)r_rzs r1ra_RepeatSampler.__init__s r>c#N# [UR5ShvN MN7fru)iterrzrvs r1rr_RepeatSampler.__iter__s"DLL) ) ) )s %#%rN)rrrr__doc__rarrrrr>r1rrs *r>r)r):r loggingr contextlibr functoolsr itertoolsrtypingrrrr r rtorch.utils.datanumpyr constantsr r rrrdistributed_samplerrrrZrmixuprscheduled_samplerrrtransforms_factoryr getLoggerrr8r2r=r@rrXrIrrrrrrrrrrobjectrrr>r1rs$ == B7L)#O0   H %$N  i5i5X =(")-/3/3!-1 "$&&* '"7!5!$(#',0)- !&$||F3#(-#'#!&dh6::>!"%//3'++.%)aH|%99:H#uS#Xc3m0DDEHH H  H  HHHH"#HeUl+,HeUl+,HHHH $E?!H"#H$"%H&sm'H()H*+H,-H.E3J/H05#: 1H23H45H65/7H8C=9H:%SM;H<X&=H>?H@AHB;;CHD EHFGHH"&IHJ!KHLMHNOHP%XeCsCx%PSUXZ]P]J^4^._%`aQHR%Xc]3SHT'x7UHVWHX #YHZ (}[H\ %]H^$)_H`c]aHV&EKK,,77&( *V *r>