B #=6\Íã@s\ddlZddlZddlZddlmZdd„Zdd„Zdd„Zd d „Zd a ej ddd „ƒZ dS)éN)Údefault_generatorcCst |¡dS)zoSets the random number generator state. Args: new_state (torch.ByteTensor): The desired state N)rZ set_state)Z new_state©rú//tmp/pip-install-l3r2oljg/torch/torch/random.pyÚ set_rng_statesrcCst ¡S)zBReturns the random number generator state as a `torch.ByteTensor`.)rZ get_staterrrrÚ get_rng_statesrcCs.t|ƒ}ddl}|jjs$|j |¡t |¡S)zSets the seed for generating random numbers. Returns a `torch._C.Generator` object. Args: seed (int): The desired seed. rN)ÚintÚ torch.cudaÚcudaZ _in_bad_forkZmanual_seed_allrÚ manual_seed)ÚseedÚtorchrrrr s  r cCst ¡S)zSReturns the initial seed for generating random numbers as a Python `long`. )rÚ initial_seedrrrrr &sr FTÚfork_rngÚdevicesc csþddl}|sdVdS|dkr\|j ¡}|dkrNtsNt dj|||d¡datt|ƒƒ}nt|ƒ}|  ¡}g}x4|D],}|j  |¡|  |j  ¡¡WdQRXqvWz dVWd|  |¡x:t ||ƒD],\}} |j  |¡|j  | ¡WdQRXqÈWXdS)aØ Forks the RNG, so that when you return, the RNG is reset to the state that it was previously in. Arguments: devices (iterable of CUDA IDs): CUDA devices for which to fork the RNG. CPU RNG state is always forked. By default, :meth:`fork_rng` operates on all devices, but will emit a warning if your machine has a lot of devices, since this function will run very slowly in that case. If you explicitly specify devices, this warning will be supressed enabled (bool): if ``False``, the RNG is not forked. This is a convenience argument for easily disabling the context manager without having to reindent your Python code. rNéaôCUDA reports that you have {num_devices} available devices, and you have used {caller} without explicitly specifying which devices are being used. For safety, we initialize *every* CUDA device by default, which can be quite slow if you have a lot of GPUs. If you know that you are only making use of a few CUDA devices, set the environment variable CUDA_VISIBLE_DEVICES or the '{devices_kw}' keyword argument of {caller} with the set of devices you are actually using. For example, if you are using CPU only, set CUDA_VISIBLE_DEVICES= or devices=[]; if you are using GPU 0 only, set CUDA_VISIBLE_DEVICES=0 or devices=[0]. To initialize all devices and suppress this warning, set the '{devices_kw}' keyword argument to `range(torch.cuda.device_count())`.)Ú num_devicesZcallerZ devices_kwT)rr Z device_countÚ_fork_rng_warned_alreadyÚwarningsÚwarnÚformatÚlistÚrangerÚdeviceÚappendrÚzip) rZenabledZ_callerZ _devices_kwr rZ cpu_rng_stateZgpu_rng_statesrZ gpu_rng_staterrrr0s0      )NTrr) r Ú contextlibrZtorch._Crrrr r rÚcontextmanagerrrrrrÚs