o i9:j+%@sddlZddlZddlmZddlmZddlZgdZer$ddlm Z ddl m Z dej ddfd d Z dej fd d Zdejjfd dZdejjfddZdefddZdddZdefddZdaej     d defddZdejdBfddZdS)!N) Generator) TYPE_CHECKING) set_rng_state get_rng_state manual_seedseed initial_seedfork_rngthread_safe_generator) WorkerInfo)default_generator new_statereturncCst|dS)zSets the random number generator state. .. note:: This function only works for CPU. For CUDA, please use :func:`torch.manual_seed`, which works for both CPU and CUDA. Args: new_state (torch.ByteTensor): The desired state N)r set_state)r rS/home/nk/hobo-godmode/plappi-mvp/.venv/lib/python3.10/site-packages/torch/random.pyrs rcCtS)zReturns the random number generator state as a `torch.ByteTensor`. .. note:: The returned state is for the default generator on CPU only. See also: :func:`torch.random.fork_rng`. )r get_staterrrrr'srcCst|S)aSets the seed for generating random numbers on all devices. Returns a `torch.Generator` object. Args: seed (int): The desired seed. Value must be within the inclusive range `[-0x8000_0000_0000_0000, 0xffff_ffff_ffff_ffff]`. Otherwise, a RuntimeError is raised. Negative inputs are remapped to positive values with the formula `0xffff_ffff_ffff_ffff + seed`. )_manual_seed_impl)rrrrr1s rcCst|}ddl}|js|j|ddl}|js"|j|ddl}|j s1|j |ddl }|j s@|j |t |t |S)Nr)int torch.cudacuda_is_in_bad_forkmanual_seed_all torch.mpsmpsr torch.xpuxpu torch.mtiamtia_seed_custom_devicer rtorchrrrr>s         rcCst}ddl}|js|j|ddl}|js"|j|ddl }|j s1|j |ddl }|j s@|j |t ||S)zSets the seed for generating random numbers to a non-deterministic random number on all devices. Returns a 64 bit number used to seed the RNG. rN)r rrrrrrrrrrrrr r!rrrrYs        rcCst|}tj}tt|rPtt|}d}d}t||r2t||r2t||s0t|||d Sd Sd|d}|d|d|d|d7}tj|td d d Sd S) zSets the seed to generate random numbers for custom device. Args: seed (int): The desired seed. See [Note: support the custom device with privateuse1] rrzSet seed for `z0` device does not take effect, please add API's `z` and `z` to `z` device module. stacklevelN) rr"_C_get_privateuse1_backend_namehasattrgetattrwarningswarn UserWarning)rcustom_backend_namecustom_device_mod_bad_fork_name_seed_all_namemessagerrrr ws       r cCr)zReturns the initial seed for generating random numbers as a Python `long`. .. note:: The returned seed is for the default generator on CPU only. )r rrrrrrsrFTr devicesrc #s||dkr dVdSt|j}tt|ddur$td|dd|s+dVdS|dur{}|dkrttst|d|d|d |d |d |d |d |d|d|d|d|d}tj |dddat t |}nt |}t }fdd|D}zdVWt |t||D] \} } | | qdSt |t||D] \} } | | qw)az Forks the RNG, so that when you return, the RNG is reset to the state that it was previously in. Args: devices (iterable of Device IDs): 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 suppressed enabled (bool): if ``False``, the RNG is not forked. This is a convenience argument for easily disabling the context manager without having to delete it and unindent your Python code under it. device_type (str): device type str, default is `cuda`. As for supported device, see details in :ref:`accelerator` metaNztorch has no module of `z`, you should register z,a module by `torch._register_device_module`.z reports that you have z& available devices, and you have used z_ without explicitly specifying which devices are being used. For safety, we initialize *every* zA device by default, which can be quite slow if you have a lot of z5s. If you know that you are only making use of a few z' devices, set the environment variable z_VISIBLE_DEVICES or the 'z' keyword argument of z with the set of devices you are actually using. For example, if you are using CPU only, set device.upper()_VISIBLE_DEVICES= or devices=[]; if you are using device 0 only, set zb_VISIBLE_DEVICES=0 or devices=[0]. To initialize all devices and suppress this warning, set the 'z#' keyword argument to `range(torch.z.device_count())`.r%Tcsg|]}|qSr)r).0device device_modrr szfork_rng..)r"r8typer* RuntimeError device_count_fork_rng_warned_alreadyupperr+r,listrangerrzip) r3enabled_caller _devices_kw device_type num_devicesr2 cpu_rng_statedevice_rng_statesr8device_rng_staterr9rr sp          cCs:ddlm}|}|dur|jdkr|jdur|jjSdS)asReturns a thread-safe random number generator for use in DataLoader workers. This function provides a convenient way for transforms and user code to use thread-safe random number generation without manually checking worker context. When called in a DataLoader thread worker, returns the worker's thread-local :class:`torch.Generator`. When called in the main process or process workers, returns ``None`` (which causes PyTorch functions to use the default global RNG). Returns: Optional[torch.Generator]: Thread-local generator in thread workers, None otherwise. Example:: >>> from torch.random import thread_safe_generator >>> generator = thread_safe_generator() >>> torch.randint(0, 10, (5,), generator=generator) Example with transforms:: >>> from torch.random import thread_safe_generator >>> class MyRandomTransform: ... def __call__(self, img): ... generator = thread_safe_generator() ... offset = torch.randint(0, 10, (2,), generator=generator) ... return img[..., offset[0]:, offset[1]:] r)get_worker_infoNthread)torch.utils.datarL worker_methodrngtorch_generator)rL worker_inforrrr s   r )rN)NTr r3r) contextlibr+collections.abcrtypingrr"__all__torch.utils.data._utils.workerr torch._Cr Tensorrrr'rrrrr rr?contextmanagerr r rrrrs6          P