tj4ddlZddlmZmZddlmZddlmZmZddl Z ddl m Z gdZ GddZGdd eZGd d eZGd d ZdS)N)ABCabstractmethod) TracebackType)Any NamedTuple)JoinHookJoinableJoinc*eZdZdZddZdeddfdZdS)ra This defines a join hook, which provides two entry points in the join context manager. Entry points : a main hook, which is called repeatedly while there exists a non-joined process, and a post-hook, which is called once all processes have joined. To implement a join hook for the generic join context manager, define a class that inherits from :class:`JoinHook` and override ``main_hook()`` and ``post_hook()`` as appropriate. returnNcdS)zCall this hook while there exists a non-joined process to shadow collective communications in a training iteration. Training iteration i.e., in one forward pass, backward pass, and optimizer step. Nselfs _/var/www/html/CA-Chatbot/venv/lib/python3.11/site-packages/torch/distributed/algorithms/join.py main_hookzJoinHook.main_hookis_last_joinercdS)aK Call hook after all processes have joined. It is passed an additional ``bool`` argument ``is_last_joiner``, which indicates if the rank is one of the last to join. Arguments: is_last_joiner (bool): ``True`` if the rank is one of the last to join; ``False`` otherwise. Nr)rrs r post_hookzJoinHook.post_hook rrr N)__name__ __module__ __qualname____doc__rboolrrrrrrsT                rrceZdZdZedfd ZedefdZeede j fdZ eede fdZ xZS) r a_ This defines an abstract base class for joinable classes. A joinable class (inheriting from :class:`Joinable`) should implement :meth:`join_hook`, which returns a :class:`JoinHook` instance, in addition to :meth:`join_device` and :meth:`join_process_group` that return device and process group information, respectively. r Nctt|_dSN)super__init__ _JoinConfigconstruct_disabled_join_config _join_config)r __class__s rr"zJoinable.__init__7s3 'FFHHrc dS)a Return a :class:`JoinHook` instance for the given :class:`Joinable`. Arguments: kwargs (dict): a :class:`dict` containing any keyword arguments to modify the behavior of the join hook at run time; all :class:`Joinable` instances sharing the same join context manager are forwarded the same value for ``kwargs``. Nr)rkwargss r join_hookzJoinable.join_hook<s  rcdS)zeReturn the device from which to perform collective communications needed by the join context manager.Nrrs r join_devicezJoinable.join_deviceI  rcdS)zfReturns the process group for the collective communications needed by the join context manager itself.Nrrs rjoin_process_groupzJoinable.join_process_groupOr,rr)rrrrrr"rr)propertytorchdevicer+rr. __classcell__)r&s@rr r ,sIIIII^I  X    ^   U\   ^X  C   ^X     rr cHeZdZUdZeed<eed<eed<edZdS)r#zdThis includes all fields needed from a :class:`Joinable` instance for the join context manager side.enablethrow_on_early_terminationis_first_joinablec&tdddS)zReturn a :class:`_JoinConfig` instance indicating that join-related logic should be disabled. e.g. if the caller is not in a join context manager. Fr4r5r6)r#rrrr$z*_JoinConfig.construct_disabled_join_config]s" Ue    rN)rrrrr__annotations__ staticmethodr$rrrr#r#VsVoo LLL $$$$  \   rr#ceZdZdZ ddeededefdZdd Zdd Z d Z d e e d zde d zde d zfdZdZdZedefdZd S)r a This class defines the generic join context manager, which allows custom hooks to be called after a process joins. These hooks should shadow the collective communications of non-joined processes to prevent hanging and erroring and to ensure algorithmic correctness. Refer to :class:`JoinHook` for details about the hook definition. .. warning:: The context manager requires each participating :class:`Joinable` to call the method :meth:`notify_join_context()` before its own per- iteration collective communications to ensure correctness. .. warning:: The context manager requires that all ``process_group`` attributes in the :class:`JoinHook` objects are the same. If there are multiple :class:`JoinHook` objects, then the ``device`` of the first is used. The process group and device information is used for checking for non- joined processes and for notifying processes to throw an exception if ``throw_on_early_termination`` is enabled, both of which using an all- reduce. Arguments: joinables (List[Joinable]): a list of the participating :class:`Joinable` s; their hooks are iterated over in the given order. enable (bool): a flag enabling uneven input detection; setting to ``False`` disables the context manager's functionality and should only be set when the user knows the inputs will not be uneven (default: ``True``). throw_on_early_termination (bool): a flag controlling whether to throw an exception upon detecting uneven inputs (default: ``False``). Example:: >>> import os >>> import torch >>> import torch.distributed as dist >>> import torch.multiprocessing as mp >>> # xdoctest: +SKIP >>> import torch.nn.parallel.DistributedDataParallel as DDP >>> import torch.distributed.optim.ZeroRedundancyOptimizer as ZeRO >>> from torch.distributed.algorithms.join import Join >>> >>> # On each spawned worker >>> def worker(rank): >>> dist.init_process_group("nccl", rank=rank, world_size=2) >>> model = DDP(torch.nn.Linear(1, 1).to(rank), device_ids=[rank]) >>> optim = ZeRO(model.parameters(), torch.optim.Adam, lr=0.01) >>> # Rank 1 gets one more input than rank 0 >>> inputs = [torch.tensor([1.]).to(rank) for _ in range(10 + rank)] >>> with Join([model, optim]): >>> for input in inputs: >>> loss = model(input).sum() >>> loss.backward() >>> optim.step() >>> # All ranks reach here without hanging/erroring TF joinablesr4r5c t|dkrtd||_fd|jD|_||_||_||dS)Nrz7The join context manager requires at least one joinablec*g|]}|jdiS)r)r)).0joinabler(s r z!Join.__init__..s:   -5 H  ( ( ( (   r)len ValueError _joinables _join_hooks_enable_throw_on_early_termination_set_joinable_configs_extract_dist_info)rr<r4r5r(s `rr"z Join.__init__s y>>Q  VWW W#    9=    +E( ""$$$ !!!!!rr Nct|jdkrtd}|jD]%}t|j|j||_d}&dS)zESet the :class:`_JoinConfig` of each participating :class:`Joinable`.rTr8FN)rBrDAssertionErrorr#rFrGr%)rr6r@s rrHzJoin._set_joinable_configssn t  1 $ $   & &H$/|+/+K"3%%%H ! !&    & &rcd}d}|jD]/}||j}n||jkrtd||j}0||_t j|j|_||_dS)a Extract the process group and device information from the joinables. If there are multiple joinables, then the context manager uses the first specified device. Preconditions: ``self._joinables`` is not ``None`` and is non-empty. Raises: ValueError If there are multiple conflicting ``process_group`` attributes among the ``Joinable`` objects. Nz7Using join context manager with multiple process groups) rDr.rCr+_process_groupdistget_rank_rank_device)r process_groupr1r@s rrIzJoin._extract_dist_infos  . .H$ ( ; ("=== M~!-+]4#677  rcdSr rrs r __enter__zJoin.__enter__rrtypevalue tracebackc |jr|rdSd}d}d}d}tjd|s||kr%tjd|d|jd |d d |}|dkrd}n@|jr||jD]} | d}|d z }||jD]} | |dS)z Repeatedly runs the main hooks until all processes join; then, runs the post-hooks. Raises: RuntimeError If ``throw_on_early_termination=True``. NFTrioncez+Detected uneven input skew of greater than z. This means that rank z has at least zz fewer inputs than other currently-active ranks. This level of skew could lead to performance degradation during training.) stacklevel) rFwarnings simplefilterwarnrP_get_num_nonjoined_procsrG_notify_procs_to_terminaterErr) rrUrVrWall_procs_joinedriWARN_THRESHOLDnum_nonjoined_procsr)s r__exit__z Join.__exit__s_| t  F  f%%%" >!! 3%33z331?333 !#'"?"?"A"A "a''#'  3633555"&!1**I''))))!&Q1# 6) 0 0I    / / / / 0 0rctjd|j}tj||j|S)zaReturn the number of non-joined processes by shadowing an all-reduce in the non-joined processes.r\r1group)r0zerosrQrN all_reducerMitem)rres rr`zJoin._get_num_nonjoined_procssD#k!DLAAA +43FGGGG"'')))rctjd|j}tj||jt d|jd)zSchedule an all-reduce to notify non-joined processes to terminate. Also raise a ``RuntimeError`` indicating that the current process has exhausted its inputs. r\rhrizRank z exhausted all inputs.)r0onesrQrNrlrM RuntimeErrorrP)rros rrazJoin._notify_procs_to_terminate!sN z!DL111 D$78888E4:EEEFFFrr@ct|ds tdt|d|j}|jr|jsdS|j}|j}tj d|}tj ||d}|j rQtj d|}tj || |}|rtd |S) aH Notifies the join context manager that the calling process has not yet joined. Then, if ``throw_on_early_termination=True``, checks if uneven inputs have been detected (i.e. if one process has already joined) and throws an exception if so. This method should be called from a :class:`Joinable` object before its per-iteration collective communications. For example, this should be called at the beginning of the forward pass in :class:`DistributedDataParallel`. Only the first :class:`Joinable` object passed into the context manager performs the collective communications in this method, and for the others, this method is vacuous. Arguments: joinable (Joinable): the :class:`Joinable` object calling this method. Returns: An async work handle for the all-reduce meant to notify the context manager that the process has not yet joined if ``joinable`` is the first one passed into the context manager; ``None`` otherwise. r%zCheck that the z/ constructor calls the ``Joinable`` constructorNr\rhT)rjasync_oprizLDetected at least one rank that exhausted inputs. Throwing across all ranks.)hasattrrKrUr%r6r4r+r.r0rorNrlr5rkrmrp)r@ join_configr1rRroworkrk should_throws rnotify_join_contextzJoin.notify_join_context*s 4x00  +$x..+++  + , K4F 4% 3 z!F+++t=4HHH  1 K&111E OE 7 7 7 7 ::<""%) """"$ & & & &>30=!D(30t#30!4' 30303030j*** GGG5h555\555rr )r]abcrrtypesrtypingrrr0torch.distributed distributedrN__all__rr r#r rrrrs1########""""""""  + * *        <' ' ' ' ' s' ' ' T     *   $xxxxxxxxxxr