a PSic@sfddlZddlZddlmZmZddlmZd eeeeedddZ ej dGd d d ej Z dS) N)nnTensor)_log_api_usage_onceT)inputpmodetrainingreturncCstjstjstt|dks,|dkr:td||dvrPtd||r\|dkr`|Sd|}|dkr|jdgdg|jd}n dg|j}tj ||j |j d }| |}|dkr| |||S) a Implements the Stochastic Depth from `"Deep Networks with Stochastic Depth" `_ used for randomly dropping residual branches of residual architectures. Args: input (Tensor[N, ...]): The input tensor or arbitrary dimensions with the first one being its batch i.e. a batch with ``N`` rows. p (float): probability of the input to be zeroed. mode (str): ``"batch"`` or ``"row"``. ``"batch"`` randomly zeroes the entire input, ``"row"`` zeroes randomly selected rows from the batch. training: apply stochastic depth if is ``True``. Default: ``True`` Returns: Tensor[N, ...]: The randomly zeroed tensor. gg?z4drop probability has to be between 0 and 1, but got )batchrowz0mode has to be either 'batch' or 'row', but got r r)dtypedevice)torchjit is_scripting is_tracingrstochastic_depth ValueErrorshapendimemptyrr bernoulli_div_)rrrr Z survival_ratesizenoiser\/var/www/html/django/DPS/env/lib/python3.9/site-packages/torchvision/ops/stochastic_depth.pyrs"    rcsHeZdZdZeeddfdd ZeedddZed d d Z Z S) StochasticDepthz' See :func:`stochastic_depth`. N)rrr cs"tt|||_||_dSN)super__init__rrr)selfrr __class__rrr"7s zStochasticDepth.__init__)rr cCst||j|j|jSr )rrrr )r#rrrrforward=szStochasticDepth.forward)r cCs"|jjd|jd|jd}|S)Nz(p=z, mode=))r%__name__rr)r#srrr__repr__@szStochasticDepth.__repr__) r( __module__ __qualname____doc__floatstrr"rr&r* __classcell__rrr$rr2sr)T)rtorch.fxrrutilsrr.r/boolrfxwrapModulerrrrrs  '