§
    ·¯¸jøU  ã            b       ó\  — d Z ddlZddlZddlmZ ddlmZ ddlmZ ddl	m
Z
mZmZmZmZ ddlZddlZddlZddlmZmZ dd	lmZmZ dd
lmZmZ ddlmZ ddlmZ ddl m!Z!m"Z" ddl#m$Z$  ej%        e&¦  «        Z'd„ Z(d„ Z) G d„ d¦  «        Z*dVd„Z+dddddddddddddddddddeeddddddddej,         ej-        d¦  «        ddddddddddddddf-d eeef         d!ee.ee.e.f         ee.e.e.f         f         d"e.d#e/d$e/d%e0d&e1d'e.d(e/d)ee1         d*eee0e0f                  d+eee0e0f                  d,e0d-e0d.e0d/ee0         d0e0d1e0d2ee1         d3e.d4e.d5e1d6ee0d7f         d8ee0d7f         d9e.d:e/d;ee0         d<ee1         d=ee.         d>ee
         d?e/d@e/dAej2        dBej-        dCe/dDe/dEe/dFe1dGe/dHeeee.ee.e.f         ee.e.e.f         f                           dIeee.                  dJeee0                  dKe.dLe1dMee.         dNe0dOe0dPee.         f`dQ„Z3 G dR„ dSej4        j5        j6        ¦  «        Z7 G dT„ dUe8¦  «        Z9dS )Wa   Loader 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
é    N)Úsuppress)Úpartial)Úrepeat)ÚCallableÚOptionalÚSequenceÚTupleÚUnioné   )ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STD)ÚIterableImageDatasetÚImageDataset)ÚOrderedDistributedSamplerÚRepeatAugSampler)ÚRandomErasing)ÚFastCollateMixup)ÚScheduledBatchSamplerÚScheduledTransformDataset©Úcreate_transformc           	      óØ  — t          | d         t          ¦  «        sJ ‚t          | ¦  «        }t          | d         d         t          ¦  «        �rft          | d         d         t          j        ¦  «        }t          | d         d         ¦  «        }||z  }t          j        |t
          j        ¬¦  «        }t          j        |g| d         d         d         j        ¢R t
          j	        ¬¦  «        }t          |¦  «        D ]±}t          | |         d         ¦  «        |k    sJ ‚t          |¦  «        D ]~}| |         d         ||||z  z   <   |r;||||z  z   xx         t          j        | |         d         |         ¦  «        z  cc<   ŒV||||z  z   xx         | |         d         |         z  cc<   ŒŒ²||fS t          | d         d         t          j        ¦  «        r¸t          j        d„ | D ¦   «         t
          j        ¬¦  «        }t          |¦  «        |k    sJ ‚t          j        |g| d         d         j        ¢R t
          j	        ¬¦  «        }t          |¦  «        D ]0}||xx         t          j        | |         d         ¦  «        z  cc<   Œ1||fS t          | d         d         t
          j        ¦  «        r±t          j        d„ | D ¦   «         t
          j        ¬¦  «        }t          |¦  «        |k    sJ ‚t          j        |g| d         d         j        ¢R t
          j	        ¬¦  «        }t          |¦  «        D ])}||                              | |         d         ¦  «         Œ*||fS J ‚)zd A fast collation function optimized for uint8 images (np array or torch) and int64 targets (labels)r   )Údtyper   c                 ó   — g | ]
}|d          ‘ŒS ©r   © ©Ú.0Úbs     ú`/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/data/loader.pyú
<listcomp>z fast_collate.<locals>.<listcomp>4   ó   € Ð4Ð4Ð4¨  !¤Ð4Ð4Ð4ó    c                 ó   — g | ]
}|d          ‘ŒS r   r   r   s     r    r!   z fast_collate.<locals>.<listcomp>;   r"   r#   )Ú
isinstanceÚtupleÚlenÚnpÚndarrayÚtorchÚzerosÚint64ÚshapeÚuint8ÚrangeÚ
from_numpyÚtensorÚTensorÚcopy_)	ÚbatchÚ
batch_sizeÚis_npÚinner_tuple_sizeÚflattened_batch_sizeÚtargetsr1   ÚiÚjs	            r    Úfast_collater<      sa  € å�e˜A”h¥Ñ&Ô&Ð&Ð&Ð&Ý�U‘”€JÝ�%˜”(˜1”+�uÑ%Ô%ñ  õ ˜5 œ8 Aœ;­¬
Ñ3Ô3ˆÝ˜u Qœx¨œ{Ñ+Ô+ÐØ)Ð,<Ñ<ÐÝ”+Ð2½%¼+ÐFÑFÔFˆÝ”Ð2ÐJ°U¸1´X¸a´[À´^Ô5IÐJÐJÕRWÔR]Ð^Ñ^Ô^ˆÝ�zÑ"Ô"ð 	Að 	AˆAÝ�u˜Q”x ”{Ñ#Ô#Ð'7Ò7Ð7Ð7Ð7ÝÐ+Ñ,Ô,ð Að A�Ø.3°A¬h°q¬k�˜˜A 
™NÑ*Ñ+Øð AØ˜1˜q :™~Ñ-Ð.Ð.Ô.µ%Ô2BÀ5ÈÄ8ÈAÄ;ÈqÄ>Ñ2RÔ2RÑRÐ.Ð.Ñ.Ð.à˜1˜q :™~Ñ-Ð.Ð.Ô.°%¸´(¸1´+¸a´.Ñ@Ð.Ð.Ñ.Ð.ðAð �wˆÐÝ	�E˜!”H˜Q”K¥¤Ñ	,Ô	,ð Ý”,Ð4Ð4¨eÐ4Ñ4Ô4½E¼KÐHÑHÔHˆÝ�7‰|Œ|˜zÒ)Ð)Ð)Ð)Ý”˜jÐ=¨5°¬8°A¬;Ô+<Ð=Ð=ÅUÄ[ÐQÑQÔQˆÝ�zÑ"Ô"ð 	7ð 	7ˆAØ�1ˆIˆIŒI�Ô)¨%°¬(°1¬+Ñ6Ô6Ñ6ˆIˆI‰IˆIØ�wˆÐÝ	�E˜!”H˜Q”K¥¤Ñ	.Ô	.ð Ý”,Ð4Ð4¨eÐ4Ñ4Ô4½E¼KÐHÑHÔHˆÝ�7‰|Œ|˜zÒ)Ð)Ð)Ð)Ý”˜jÐ=¨5°¬8°A¬;Ô+<Ð=Ð=ÅUÄ[ÐQÑQÔQˆÝ�zÑ"Ô"ð 	)ð 	)ˆAØ�1ŒI�OŠO˜E !œH QœKÑ(Ô(Ð(Ð(Ø�wˆÐàˆur#   c                 óp  — t          | t          t          f¦  «        st          t          | |¦  «        ¦  «        } n{t	          | ¦  «        |k    rKt          j        | ¦  «                             ¦   «         }|f|z  } t           	                    d| › d�¦  «         nt	          | ¦  «        |k    s
J d¦   «         ‚| S )Nz@Pretrained mean/std different shape than model, using avg value ú.z-normalization stats must match image channels)
r%   r&   Úlistr   r'   r(   ÚmeanÚitemÚ_loggerÚwarning)ÚxÚnÚx_means      r    Úadapt_to_chsrG   E   s¢   € Ý�a�%¥˜Ñ'Ô'ð LÝ•&˜˜A‘,”,ÑÔˆˆÝ	ˆQ‰Œ�1ŠˆÝ”˜‘”—’Ñ"Ô"ˆØˆI˜‰MˆÝ�ŠÐ_Ð[\Ð_Ð_Ð_Ñ`Ô`Ð`Ð`å�1‰vŒv˜Š{ˆ{ˆ{ÐK‰{Œ{ˆ{Ø€Hr#   c                   óZ  — e Zd ZdZeed ej        d¦  «        dddddd	f
d
ej        j	        j
        deedf         deedf         dedej        deej                 dededededefd„Zd„ Zd„ Zed„ ¦   «         Zed„ ¦   «         Zed„ ¦   «         Zed„ ¦   «         Zej        d„ ¦   «         ZdS )ÚPrefetchLoaderzýPrefetch batches to the device on a side stream and normalize them there.

    The transfer of the next batch overlaps the consumer's work on the current one. Batches must
    be consumed on the stream that is current when the iterator is advanced.
    é   ÚcudaNFç        Úconstr   r   Úloaderr@   .ÚstdÚchannelsÚdeviceÚ	img_dtypeÚfp16Úre_probÚre_modeÚre_countÚre_num_splitsc                 ó¨  — t          ||¦  «        }t          ||¦  «        }d|ddf}|| _        || _        |rt          j        }|pt          j        | _        t          j        d„ |D ¦   «         ||¬¦  «                             |¦  «        | _	        t          j        d„ |D ¦   «         ||¬¦  «                             |¦  «        | _
        |dk    rt          ||	|
||¬¦  «        | _        nd | _        |j        dk    ot          j                             ¦   «         | _        |j        dk    ot          j                             ¦   «         | _        d | _        d S )	Nr   c                 ó   — g | ]}|d z  ‘ŒS ©éÿ   r   ©r   rD   s     r    r!   z+PrefetchLoader.__init__.<locals>.<listcomp>q   s   € Ð#Ð#Ð#˜ˆQ�‰WÐ#Ð#Ð#r#   )rQ   r   c                 ó   — g | ]}|d z  ‘ŒS rZ   r   r\   s     r    r!   z+PrefetchLoader.__init__.<locals>.<listcomp>s   s   € Ð"Ð"Ð"˜ˆQ�‰WÐ"Ð"Ð"r#   rL   )ÚprobabilityÚmodeÚ	max_countÚ
num_splitsrQ   rK   Únpu)rG   rN   rQ   r*   Úfloat16Úfloat32rR   r1   Úviewr@   rO   r   Úrandom_erasingÚtyperK   Úis_availableÚis_cudarb   Úis_npuÚstream)ÚselfrN   r@   rO   rP   rQ   rR   rS   rT   rU   rV   rW   Únormalization_shapes                r    Ú__init__zPrefetchLoader.__init__X   sg  € õ ˜D (Ñ+Ô+ˆÝ˜3 Ñ)Ô)ˆØ  (¨A¨qÐ1ÐàˆŒØˆŒØð 	&åœˆIØ"Ð3¥e¤mˆŒÝ”LØ#Ð#˜dÐ#Ñ#Ô#¨F¸)ðEñ Eô EßEIÂTÐJ]ÑE^ÔE^ð 	Œ	å”<Ø"Ð"˜cÐ"Ñ"Ô"¨6¸ðDñ Dô DßDHÂDÐI\ÑD]ÔD]ð 	Œà�RŠ<ˆ<Ý"/Ø#ØØ"Ø(Øð#ñ #ô #ˆDÔÐð #'ˆDÔØ”{ fÒ,ÐJµ´×1HÒ1HÑ1JÔ1JˆŒØ”k UÒ*ÐG­u¬y×/EÒ/EÑ/GÔ/GˆŒØˆŒˆˆr#   c              #   ó¨  K  — d}d }| j         rt          j        }n| j        rt          j        }|�^| j        € |                     | j        ¬¦  «        | _        t          |j        | j        ¬¦  «        }| 	                    | j        ¬¦  «        }n	t          }d }| j        D �](\  }}|�| j                             |¦  «          |¦   «         5  |                     | j        d¬¦  «        }|                     | j        d¬¦  «        }|                     | j        ¦  «                             | j        ¦  «                             | j        ¦  «        }| j        �|                      |¦  «        }d d d ¦  «         n# 1 swxY w Y   |s||fV — nd}|�5| 	                    | j        ¬¦  «        }|                     | j        ¦  «         |}|}�Œ*|s||fV — d S d S )NT)rQ   )rk   )rQ   Únon_blockingF)ri   r*   rK   rj   rb   rk   ÚStreamrQ   r   Úcurrent_streamr   rN   Úwait_streamÚtorR   Úsub_r@   Údiv_rO   rf   )	rl   ÚfirstÚ
device_modÚstream_contextÚconsumer_streamÚ
next_inputÚnext_targetÚinputÚtargets	            r    Ú__iter__zPrefetchLoader.__iter__‚   sJ  è è € Øˆàˆ
ØŒ<ð 	#ÝœˆJˆJØŒ[ð 	#ÝœˆJàÐ!ØŒ{Ð"ð )×/Ò/°t´{Ð/ÑCÔC�”Ý$ ZÔ%6¸t¼{ÐKÑKÔKˆNØ(×7Ò7¸t¼{Ð7ÑKÔKˆOˆOå%ˆNØ"ˆOà'+¤{ð 	!ñ 	!Ñ#ˆJ˜ØÐ*ð ”×'Ò'¨Ñ8Ô8Ð8à�Ñ!Ô!ð Að AØ'Ÿ]š]°$´+ÈD˜]ÑQÔQ�
Ø)Ÿnšn°D´KÈd˜nÑSÔS�Ø'Ÿ]š]¨4¬>Ñ:Ô:×?Ò?ÀÄ	ÑJÔJ×OÒOÐPTÔPXÑYÔY�
ØÔ&Ð2Ø!%×!4Ò!4°ZÑ!@Ô!@�JðAð Að Añ Aô Að Að Að Að Að Að Aøøøð Að Að Að Að ð Ø˜V�mÐ#Ð#Ð#Ð#à�àÐ*Ø",×";Ò";À4Ä;Ð";Ñ"OÔ"O�Ø×+Ò+¨D¬KÑ8Ô8Ð8àˆEØ ˆF‰Fàð 	 Ø˜�-ÐÐÐÐÐð	 ð 	 s   Ã
BE5Å5E9	Å<E9	c                 ó*   — t          | j        ¦  «        S ©N)r'   rN   ©rl   s    r    Ú__len__zPrefetchLoader.__len__²   s   € Ý�4”;ÑÔÐr#   c                 ó   — | j         j        S r�   )rN   Úsamplerr‚   s    r    r…   zPrefetchLoader.samplerµ   ó   € àŒ{Ô"Ð"r#   c                 ó   — | j         j        S r�   )rN   Úbatch_samplerr‚   s    r    rˆ   zPrefetchLoader.batch_sampler¹   s   € àŒ{Ô(Ð(r#   c                 ó   — | j         j        S r�   )rN   Údatasetr‚   s    r    rŠ   zPrefetchLoader.dataset½   r†   r#   c                 óf   — t          | j        j        t          ¦  «        r| j        j        j        S dS )NF©r%   rN   Ú
collate_fnr   Úmixup_enabledr‚   s    r    rŽ   zPrefetchLoader.mixup_enabledÁ   s-   € å�d”kÔ,Õ.>Ñ?Ô?ð 	Ø”;Ô)Ô7Ð7à�5r#   c                 ój   — t          | j        j        t          ¦  «        r|| j        j        _        d S d S r�   rŒ   )rl   rD   s     r    rŽ   zPrefetchLoader.mixup_enabledÈ   s7   € å�d”kÔ,Õ.>Ñ?Ô?ð 	5Ø34ˆDŒKÔ"Ô0Ð0Ð0ð	5ð 	5r#   )Ú__name__Ú
__module__Ú__qualname__Ú__doc__r   r   r*   rQ   ÚutilsÚdataÚ
DataLoaderr	   ÚfloatÚintr   r   ÚboolÚstrrn   r   rƒ   Úpropertyr…   rˆ   rŠ   rŽ   Úsetterr   r#   r    rI   rI   Q   s�  € € € € € ðð ð '<Ø%9ØØ#/ 5¤<°Ñ#7Ô#7Ø/3ØØØ"ØØ!"ð(ð (à”KÔ$Ô/ð(ð ˜˜s˜
Ô#ð(ð �u˜c�zÔ"ð	(ð
 ð(ð ”Lð(ð   ¤Ô,ð(ð ð(ð ð(ð ð(ð ð(ð ð(ð (ð (ð (ðT. ð . ð . ð` ð  ð  ð ð#ð #ñ „Xð#ð ð)ð )ñ „Xð)ð ð#ð #ñ „Xð#ð ðð ñ „Xðð Ôð5ð 5ñ Ôð5ð 5ð 5r#   rI   Úallc                 ó¨  — t           j        j                             ¦   «         }|j        | k    sJ ‚t          |t          ¦  «        rW ||¦  «        }t          j        |¦  «         t          j	        |¦  «         t          j                             |dz  ¦  «         d S |dv sJ ‚|dk    r)t          j                             |j        dz  ¦  «         d S d S )Nl   ÿÿ )r�   Úpartr�   )r*   r”   r•   Úget_worker_infoÚidr%   r   ÚrandomÚseedÚmanual_seedr(   )Ú	worker_idÚworker_seedingÚworker_infor£   s       r    Ú_worker_initr¨   Î   sÔ   € Ý”+Ô"×2Ò2Ñ4Ô4€KØŒ>˜YÒ&Ð&Ð&Ð&Ý�.¥(Ñ+Ô+ð 
=Øˆ~˜kÑ*Ô*ˆÝŒ�DÑÔÐÝÔ˜$ÑÔÐÝ
Œ	�Š�t˜{Ñ+Ñ,Ô,Ð,Ð,Ð,à Ð0Ð0Ð0Ð0ð ˜UÒ"Ð"ÝŒI�NŠN˜;Ô+¨{Ñ;Ñ<Ô<Ð<Ð<Ð<ð #Ð"r#   FrL   rM   g      à?gš™™™™™Ù?ÚbilinearrK   TÚconstantgÍÌÌÌÌÌä?gš™™™™™¹?rŠ   Ú
input_sizer5   Úis_trainingÚno_augrT   rU   rV   Úre_splitÚtrain_crop_modeÚscaleÚratioÚhflipÚvflipÚcolor_jitterÚcolor_jitter_probÚgrayscale_probÚgaussian_blur_probÚauto_augmentÚnum_aug_repeatsÚnum_aug_splitsÚinterpolationr@   .rO   Únum_workersÚdistributedÚcrop_pctÚ	crop_modeÚcrop_border_pixelsr�   Ú
pin_memoryrS   rR   rQ   Úuse_prefetcherÚuse_multi_epochs_loaderÚpersistent_workersr¦   Útf_preprocessingÚinput_size_choicesÚbatch_size_choicesÚbatch_choice_weightsÚbatch_choice_seedÚbatch_choice_scheduleÚbatch_schedule_epochsÚbatch_schedule_spreadÚbatch_schedule_random_mixÚnum_batchesc0                 óh	  ‡<— d}0|r|pd}0t          d5i d|“d|“d|	“d|
“d|“d|“d	|“d
|“d|“d|“d|“d|“d|“d|“d|“d|“d|“d|“d|“d|“d|“d|0“d|&“d|"“d|dk    “ŽŠ<t          |t          t          f¦  «        rt	          |¦  «        dk    r|d         nt	          |¦  «        }1|'du}2|2�rÝ|st          d¦  «        ‚|#rt          d¦  «        ‚t          | t          j        j        j	        ¦  «        rt          d ¦  «        ‚|dk    rt          d!¦  «        ‚|'st          d"¦  «        ‚g }3|'D ]Á}4t          |4t          ¦  «        r|1|4|4f}4ngt	          |4¦  «        dk    r|1g|4¢R }4nMt	          |4¦  «        dk    r+t          |4¦  «        }4|4d         |1k    rt          d#¦  «        ‚nt          d$¦  «        ‚t          d%„ |4D ¦   «         ¦  «        rt          d&¦  «        ‚|3                     |4¦  «         ŒÂ|(€|gt	          |3¦  «        z  }(n/t	          |(¦  «        t	          |3¦  «        k    rt          d'¦  «        ‚|)�/t	          |)¦  «        t	          |3¦  «        k    rt          d(¦  «        ‚ˆ<fd)„|3D ¦   «         }5d| _        t          | |5¦  «        } nt!          |fi ‰<¤Ž| _        t          | t"          ¦  «        r|                      |¬*¦  «         d}6|rtt          | t          j        j        j	        ¦  «        sP|r>|rt'          | |¬+¦  «        }6nJt          j        j        j                             | ¦  «        }6n t-          | ¦  «        }6n|dk    s
J d,¦   «         ‚|2r&|6€$t          j        j                             | ¦  «        }6|€$|"rt0          nt          j        j        j        j        }t          j        j        j        }7|#rt8          }7t          |||t;          t<          |%¬-¦  «        |$o|dk    ¬.¦  «        }8|2rt?          |6|(|)|*||/|+|,|-|.¬/¦
  «
        |8d0<   nB|8                      |t          | t          j        j        j	        ¦  «         o|6du o||6|¬1¦  «         	  |7| fi |8¤Ž}9n5# t          $ r(}:|8 !                    d2¦  «          |7| fi |8¤Ž}9Y d}:~:nd}:~:ww xY w|"r"|r|s|nd3};tE          |9|||1|!|| |;|||0¬4¦  «        }9|9S )6aÀ  

    Args:
        dataset: The image dataset to load.
        input_size: Target input size (channels, height, width) tuple or size scalar.
        batch_size: Number of samples in a batch.
        is_training: Return training (random) transforms.
        no_aug: Disable augmentation for training (useful for debug).
        re_prob: Random erasing probability.
        re_mode: Random erasing fill mode.
        re_count: Number of random erasing regions.
        re_split: Control split of random erasing across batch size.
        scale: Random resize scale range (crop area, < 1.0 => 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
    r   é   r¬   r­   r¯   r°   r±   r²   r³   r´   rµ   r¶   r·   r¸   r»   r@   rO   r¾   r¿   rÀ   rT   rU   rV   rW   rÅ   rÂ   ÚseparaterJ   Nz>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.c              3   ó"   K  — | ]
}|d k    V — ŒdS )r   Nr   )r   Ú	dimensions     r    ú	<genexpr>z create_loader.<locals>.<genexpr>ˆ  s&   è è € Ð8Ð8 i�9 ’>Ð8Ð8Ð8Ð8Ð8Ð8r#   z5All 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.c                 ó*   •— g | ]}t          |fi ‰¤Ž‘ŒS r   r   )r   ÚsizeÚtransform_kwargss     €r    r!   z!create_loader.<locals>.<listcomp>“  s+   ø€ ÐbÐbÐbÀTÕ& tÐ@Ð@Ð/?Ð@Ð@ÐbÐbÐbr#   )r¼   )Únum_repeatszORepeatAugment not currently supported in non-distributed or IterableDataset use)r¦   )r¼   r�   rÁ   Úworker_init_fnrÄ   )	Úbatch_sizesÚchoice_weightsr£   Ú	drop_lastrÎ   Úchoice_scheduleÚschedule_epochsÚschedule_spreadÚschedule_random_mixrˆ   )r5   Úshuffler…   rÜ   rÄ   rL   )
r@   rO   rP   rQ   rS   rR   rT   rU   rV   rW   r   )#Údictr%   r&   r?   r'   Ú
ValueErrorr*   r”   r•   ÚIterableDatasetÚ	TypeErrorr˜   ÚanyÚappendÚ	transformr   r   r   Úset_loader_cfgr   r½   ÚDistributedSamplerr   ÚRandomSamplerr<   Ú
dataloaderÚdefault_collater–   ÚMultiEpochsDataLoaderr   r¨   r   ÚupdateÚpoprI   )=rŠ   r«   r5   r¬   r­   rT   rU   rV   r®   r¯   r°   r±   r²   r³   r´   rµ   r¶   r·   r¸   r¹   rº   r»   r@   rO   r¼   r½   r¾   r¿   rÀ   r�   rÁ   rS   rR   rQ   rÂ   rÃ   rÄ   r¦   rÅ   rÆ   rÇ   rÈ   rÉ   rÊ   rË   rÌ   rÍ   rÎ   rW   rP   Úscheduled_batchingÚresolved_input_sizesrÖ   Ú
transformsr…   Úloader_classÚloader_argsrN   ÚeÚprefetch_re_probr×   s=                                                               @r    Úcreate_loaderrø   Þ   s³  ø€ ðX €MØð ,à&Ð+¨!ˆÝð ð ð Ø�Kðàˆvðð (˜ðð ˆeð	ð
 ˆeðð ˆeðð ˆeðð "�\ðð ,Ð+ðð &�~ðð .Ð-ðð "�\ðð $�mðð ˆTðð ˆCðð  �ð!ð" �)ð#ð$ .Ð-ð%ð& �ð'ð( �ð)ð* �ð+ð, $�mð-ð. *Ð)ð/ð0 &�~ð1ð2   !Ò#Ð#ð3Ðõ: �j¥5­$ -Ñ0Ô0ð	Ý58¸±_´_ÈÒ5IÐ5Ið 	�1Œˆå�‰YŒYð ð ,°4Ð7ÐØñ 'MØð 	_ÝÐ]Ñ^Ô^Ð^Ø"ð 	cÝÐaÑbÔbÐbÝ�g�uœ{Ô/Ô?Ñ@Ô@ð 	RÝÐPÑQÔQÐQØ˜AÒÐÝÐ`ÑaÔaÐaØ!ð 	SÝÐQÑRÔRÐRà!ÐØ&ð 	.ð 	.ˆDÝ˜$¥Ñ$Ô$ð 	eØ  $¨Ð-��Ý�T‘”˜a’�Ø Ð( 4Ð(Ð(��Ý�T‘”˜a’�Ý˜T‘{”{�Ø˜”7˜hÒ&Ð&Ý$Ð%fÑgÔgÐgð 'õ !Ð!cÑdÔdÐdÝÐ8Ð8°4Ð8Ñ8Ô8Ñ8Ô8ð ZÝ Ð!XÑYÔYÐYØ ×'Ò'¨Ñ-Ô-Ð-Ð-àÐ%Ø", µÐ4HÑ0IÔ0IÑ!IÐÐÝÐ#Ñ$Ô$­Ð,@Ñ(AÔ(AÒAÐAÝÐcÑdÔdÐdØÐ+µÐ4HÑ0IÔ0IÍSÐQeÑMfÔMfÒ0fÐ0fÝÐeÑfÔfÐfàbÐbÐbÐbÐMaÐbÑbÔbˆ
Ø ˆÔÝ+¨G°ZÑ@Ô@ˆˆå,¨ZÐLÐLÐ;KÐLÐLˆÔå�'Õ/Ñ0Ô0ð 8ð 	×Ò¨;ÐÑ7Ô7Ð7à€GØð w�: g­u¬{Ô/?Ô/OÑPÔPð wØð 	9Øð SÝ*¨7ÀÐPÑPÔP��åœ+Ô*Ô6×IÒIÈ'ÑRÔR��õ 0°Ñ8Ô8ˆGˆGà !Ò#Ð#Ð#Ð%vÑ#Ô#Ð#àð :˜g˜oÝ”+Ô"×0Ò0°Ñ9Ô9ˆàÐØ%3Ðd•\�\½¼Ô9IÔ9TÔ9dˆ
å”;Ô#Ô.€LØð -Ý,ˆåØØØÝ�|¸NÐKÑKÔKØ-ÐA°+À²/ðñ ô €Kð ð 
Ý'<ØØ*Ø/Ø"Ø!Ø#Ø1Ø1Ø1Ø 9ð(
ñ (
ô (
ˆ�OÑ$Ð$ð 	×ÒØ!Ý" 7­E¬KÔ,<Ô,LÑMÔMÐMÐqÐRYÐ]aÐRaÐqÐfqØØ!ð	 	ñ 	
ô 	
ð 	
ð6Ø�˜gÐ5Ð5¨Ð5Ð5ˆˆøÝð 6ð 6ð 6Ø�ŠÐ,Ñ-Ô-Ð-Ø�˜gÐ5Ð5¨Ð5Ð5ˆˆˆˆˆˆøøøøð6øøøð ð 
Ø&1ÐH¸&ÐH˜7˜7ÀbÐÝØØØØØØØØ$ØØØ'ð
ñ 
ô 
ˆð €Ms   Ñ	Q Ñ
RÑ#RÒRc                   ó*   ‡ — e Zd Zˆ fd„Zd„ Zd„ Zˆ xZS )rî   c                 ó  •—  t          ¦   «         j        |i |¤Ž d| _        | j        €t	          | j        ¦  «        | _        nt	          | j        ¦  «        | _        d| _        t          ¦   «                              ¦   «         | _        d S )NFT)Úsuperrn   Ú_DataLoader__initializedrˆ   Ú_RepeatSamplerr…   r   Úiterator)rl   ÚargsÚkwargsÚ	__class__s      €r    rn   zMultiEpochsDataLoader.__init__ë  sy   ø€ Ø�‰ŒÔ˜$Ð) &Ð)Ð)Ð)Ø(-ˆÔ%ØÔÐ%Ý)¨$¬,Ñ7Ô7ˆDŒLˆLå!/°Ô0BÑ!CÔ!CˆDÔØ(,ˆÔ%Ý™œ×(Ò(Ñ*Ô*ˆŒˆˆr#   c                 ój   — | j         €t          | j        ¦  «        nt          | j         j        ¦  «        S r�   )rˆ   r'   r…   r‚   s    r    rƒ   zMultiEpochsDataLoader.__len__õ  s-   € Ø$(Ô$6Ð$>�s�4”<Ñ Ô Ð ÅCÈÔHZÔHbÑDcÔDcÐcr#   c              #   ót   K  — t          t          | ¦  «        ¦  «        D ]}t          | j        ¦  «        V — Œd S r�   )r/   r'   Únextrþ   )rl   r:   s     r    r   zMultiEpochsDataLoader.__iter__ø  sF   è è € Ý•s˜4‘y”yÑ!Ô!ð 	&ð 	&ˆAÝ�t”}Ñ%Ô%Ð%Ð%Ð%Ð%ð	&ð 	&r#   )r�   r‘   r’   rn   rƒ   r   Ú__classcell__)r  s   @r    rî   rî   é  sY   ø€ € € € € ð+ð +ð +ð +ð +ðdð dð dð&ð &ð &ð &ð &ð &ð &r#   rî   c                   ó   — e Zd ZdZd„ Zd„ ZdS )rý   zH Sampler that repeats forever.

    Args:
        sampler (Sampler)
    c                 ó   — || _         d S r�   )r…   )rl   r…   s     r    rn   z_RepeatSampler.__init__  s   € ØˆŒˆˆr#   c              #   ó>   K  — 	 t          | j        ¦  «        E d {V —† Œr�   )Úiterr…   r‚   s    r    r   z_RepeatSampler.__iter__  s2   è è € ð	*Ý˜DœLÑ)Ô)Ð)Ð)Ð)Ð)Ð)Ð)Ð)ð	*r#   N)r�   r‘   r’   r“   rn   r   r   r#   r    rý   rý   ý  s<   € € € € € ðð ðð ð ð*ð *ð *ð *ð *r#   rý   )r�   ):r“   Úloggingr¢   Ú
contextlibr   Ú	functoolsr   Ú	itertoolsr   Útypingr   r   r   r	   r
   r*   Útorch.utils.dataÚnumpyr(   Ú	constantsr   r   rŠ   r   r   Údistributed_samplerr   r   rf   r   Úmixupr   Úscheduled_samplerr   r   Útransforms_factoryr   Ú	getLoggerr�   rB   r<   rG   rI   r¨   rd   rQ   r˜   r™   r—   rš   r   rø   r”   r•   r–   rî   Úobjectrý   r   r#   r    ú<module>r     s»  ððð ð €€€Ø €€€Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð Ø =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =Ð =à €€€Ø Ð Ð Ð Ø Ð Ð Ð à BÐ BÐ BÐ BÐ BÐ BÐ BÐ BØ 7Ð 7Ð 7Ð 7Ð 7Ð 7Ð 7Ð 7Ø LÐ LÐ LÐ LÐ LÐ LÐ LÐ LØ )Ð )Ð )Ð )Ð )Ð )Ø #Ð #Ð #Ð #Ð #Ð #Ø OÐ OÐ OÐ OÐ OÐ OÐ OÐ OØ 0Ð 0Ð 0Ð 0Ð 0Ð 0à
ˆ'Ô
˜HÑ
%Ô
%€ð$ð $ð $ðN	ð 	ð 	ðz5ð z5ð z5ð z5ð z5ñ z5ô z5ð z5ðz=ð =ð =ð =ð( "ØØØØØØ)-Ø/3Ø/3ØØØ!Ø-1Ø "Ø$&Ø&*Ø ØØ'Ø"7Ø!5ØØ!Ø$(Ø#'Ø,0Ø)-Ø ØØ!&¤Ø+˜uœ|¨FÑ3Ô3Ø#Ø(-Ø#'Ø#Ø!&ØdhØ6:Ø:>Ø!"Ø%/Ø/3Ø'+Ø+.Ø%)ðaHð HØ�|Ð%9Ð9Ô:ðHà˜#˜u S¨# Xœ°°c¸3À°mÔ0DÐDÔEðHð ðHð ð	Hð
 ðHð ðHð ðHð ðHð ðHð " #œðHð ˜˜e U˜lÔ+Ô,ðHð ˜˜e U˜lÔ+Ô,ðHð ðHð ðHð ðHð  $ Eœ?ð!Hð" ð#Hð$ "ð%Hð& ˜s”mð'Hð( ð)Hð* ð+Hð, ð-Hð. �E˜3�JÔð/Hð0 �5˜#�:Ôð1Hð2 ð3Hð4 ð5Hð6 ˜5”/ð7Hð8 ˜C”=ð9Hð: % SœMð;Hð< ˜XÔ&ð=Hð> ð?Hð@ ðAHðB ”;ðCHðD ”ðEHðF ðGHðH "&ðIHðJ !ðKHðL ðMHðN ðOHðP % X¨e°C¸¸sÀC¸x¼È%ÐPSÐUXÐZ]ÐP]ÔJ^Ð4^Ô._Ô%`ÔaðQHðR % X¨c¤]Ô3ðSHðT ' x°¤Ô7ðUHðV ðWHðX  #ðYHðZ  (¨œ}ð[Hð\  %ð]Hð^ $)ð_Hð` ˜c”]ðaHð Hð Hð HðV&ð &ð &ð &ð &˜EœKÔ,Ô7ñ &ô &ð &ð(*ð *ð *ð *ð *�Vñ *ô *ð *ð *ð *r#   