Fi/dZddlmZddlZddlmZddlZddlZddl m Z m Z m Z ddl mZmZ d d d Zdd e j"ddf dd Zy)zlFunctions for estimating the best YOLO batch size to use a fraction of the available CUDA memory in PyTorch.) annotationsN)deepcopy) DEFAULT_CFGLOGGERcolorstr)autocast profile_opsct|5tt|j|d|cxkrdkrnn|nd||cdddS#1swYyxYw)a)Compute optimal YOLO training batch size using the autobatch() function. Args: model (torch.nn.Module): YOLO model to check batch size for. imgsz (int, optional): Image size used for training. amp (bool, optional): Use automatic mixed precision if True. batch (int | float, optional): Fraction of GPU memory to use. If -1, use default. max_num_obj (int, optional): The maximum number of objects from dataset. dataset_size (int, optional): Total number of training images. If > 0, batch size will not exceed this value. Returns: (int): Optimal batch size computed using the autobatch() function. Notes: If 0.0 < batch < 1.0, it's used as the fraction of GPU memory to use. Otherwise, a default fraction of 0.6 is used. )enabledgg?333333?)fraction max_num_obj dataset_sizeN)r autobatchrtrain)modelimgszampbatchrrs ]/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/ultralytics/utils/autobatch.pycheck_train_batch_sizersS2 #   UO ! ! # !E/C/US#%     s 7AArcptd}tj|d|d|dzdt|j j }|j dvrtj|d||Stjjjrtj|d||Sd }d tjd d jd } tjj!|} | j"|z } tjj%||z } tjj'||z } | | | zz }tj|| d| j(d| dd| dd| dd|dd | dkrgdngd}|d kDr|Dcgc] }||ks | }}|j*j-dd} |Dcgc]}tj.||||}}t1||d||}t3t5||Dcgc]Y\}\}}|rOt7|dt8t:fr6d |dcxkr| kr(nn%|d k(s||dz r|d||dz dkDr||dg[}}}}|rt5|nggf\}}t=j>||d}t9tA||z|dz |d z }d|vr+|jCd}|||k\r|tE|dz d }|dks|dkDr tj|d |d!|d"|}|d kDr tG||}t=jH||| z| z| z }tj|d#|d$| d%| |zdd&| dd'|dzd(d) |tjjKScc}wcc}wcc}}}w#tL$rH}tj|d*|d+|d"|cYd}~tjjKSd}~wwxYw#tjjKwxYw),aAutomatically estimate the best YOLO batch size to use a fraction of the available CUDA memory. Args: model (torch.nn.Module): YOLO model to compute batch size for. imgsz (int, optional): The image size used as input for the YOLO model. fraction (float, optional): The fraction of available CUDA memory to use. batch_size (int, optional): The default batch size to use if an error is detected. max_num_obj (int, optional): The maximum number of objects from dataset. dataset_size (int, optional): Total number of training images. If > 0, batch size will not exceed this value. Returns: (int): The optimal batch size. z AutoBatch: z'Computing optimal batch size for imgsz=z at dz% CUDA memory utilization.>cpumpsz4intended for CUDA devices, using default batch-size zHRequires torch.backends.cudnn.benchmark=False, using default batch-size i@zCUDA:CUDA_VISIBLE_DEVICES0rz (z) z.2fz G total, z G reserved, z G allocated, zG free)r r )r r!r"r#r @channelsr )ndevicerr!)degNizbatch=z. outside safe range, using default batch-size .zUsing batch-size z for  zG/zG (z.0fu%) ✅zerror detected: z, using default batch-size )'rrinfonext parametersr)typewarningtorchbackendscudnn benchmarkosgetenvstripcudaget_device_properties total_memorymemory_reservedmemory_allocatednameyamlgetemptyr enumeratezip isinstanceintfloatnppolyfitroundindexmaxminpolyval empty_cache Exception)rrr batch_sizerrprefixr)gbd propertiestraf batch_sizesbchimgresultsixyxyfit_xfit_ypes rrr3sP,m $F KK6(A%XX[^L\\vwx %""$ % , ,F {{n$&!UV`Uabc ~~%%&!ijtiuvw B  0#6<<>qABCA11&9J"$A ""6*R/A ##F+b0A QU A KK6(1#R 01S'1S'VWX[U\\ijkloippvwx'("f"2JKa"-CQl1BqC C  A &B!!9DEAu{{1b%/EEc5Af+V 's;'@A  6Aq1Q4#u.AaD 1 awq1u~1Aq8I1I !I  $&sBxB8 u JJue + q8|$qt+qt3 4 7? d#AKN"AE1 . q5AH NNfXVA3.\]g\hhij kA ! A|$AJJq!$q(1,1 vh/s%s!AL;MRPQRUwVYZbehZhilYmmstu   GDF 0 &!1!4PQ[P\\]^_     s\- N08N0ON5>+O)AN:D O5 O P P )P*P PP P5)r Tr r)rtorch.nn.ModulerrErboolrz int | floatrrErrEreturnrE)rrgrrErrFrPrErrErrErirE)__doc__ __future__rr6copyrnumpyrGr2ultralytics.utilsrrrultralytics.utils.torch_utilsrr rrrrrrsr"  ;;?                    J!'' P! P! P!P! P!  P!  P! P!rq