ڧg rddlZddlmcmZddlmZ d dejdejded ed e d ejf d Z dS)N)_log_api_usage_once?noneinputstargetsalphagamma reductionreturnc tjs2tjst t tj|}tj||d}||zd|z d|z zz}|d|z |zz}|dkr||zd|z d|z zz} | |z}|dkrnI|dkr| }n.|dkr| }ntd|d|S) a Loss used in RetinaNet for dense detection: https://arxiv.org/abs/1708.02002. Args: inputs (Tensor): A float tensor of arbitrary shape. The predictions for each example. targets (Tensor): A float tensor with the same shape as inputs. Stores the binary classification label for each element in inputs (0 for the negative class and 1 for the positive class). alpha (float): Weighting factor in range (0,1) to balance positive vs negative examples or -1 for ignore. Default: ``0.25``. gamma (float): Exponent of the modulating factor (1 - p_t) to balance easy vs hard examples. Default: ``2``. reduction (string): ``'none'`` | ``'mean'`` | ``'sum'`` ``'none'``: No reduction will be applied to the output. ``'mean'``: The output will be averaged. ``'sum'``: The output will be summed. Default: ``'none'``. Returns: Loss tensor with the reduction option applied. r)r rmeansumz$Invalid Value for arg 'reduction': 'z3 Supported reduction modes: 'none', 'mean', 'sum') torchjit is_scripting is_tracingrsigmoid_focal_losssigmoidF binary_cross_entropy_with_logitsrr ValueError) rrr r r pce_lossp_tlossalpha_ts V/var/www/html/ai-engine/env/lib/python3.11/site-packages/torchvision/ops/focal_loss.pyrrs*: 9 ! ! # #0EI,@,@,B,B0./// fA0FSSSG g+Q1w;/ /C q3w5( )D zz'/QY1w;$??~F f  yy{{ e  xxzz r9 r r r    K)rrr) rtorch.nn.functionalnn functionalrutilsrTensorfloatstrrr rr)s ''''''  33 L3 \3 3  3  3  \ 333333r