:j &ddlZddlZddlZddlmZddlmZmZddlm Z ddl m Z ddl m Z mZmZmZmZddlZddlZddlmZddlmZdd lmZdd lmZdd lmZddlZdd l m!Z!dd lm"Z"m#Z#m$Z$m%Z%m&Z&ddl'm(Z(ddl)m*Z*ddl+m,Z,ddl-m.Z.edZ/e/r+ddl0Z0ddl1m2Z2m3Z3m4Z4m5Z5m6Z6e6e ddl1m7Z7e7dne:e;d<e:e;d<edZ<Gddejzj|j~Z?ee?fZ@eee?fZAeeBeeBe?fZCeeeDeEeFeEe fe4fZGeegefZHeegeee?ffZIGdde3ZJGd d!e"ZKGd"d#ZLd$eBd%eEfd&ZM d9d$eBd'eee4eFeEe ffd%eFeEe ffd(ZNd$eBd%ee4geFeEe fffd)ZOd*ee ePe d+ffd,eFeEe fd%e fd-ZQd.e:d%eeEfd/ZRd0eeeBfd%eBfd1ZSd2eFd3eEd0e d%dfd4ZTGd5d6ZUd$eBe<d7eFeEe fd%e=4.27.7)ActionConfigFileArgumentParser Namespaceclass_from_function"register_unresolvable_import_paths)set_parsing_settingsT)config_read_mode_fsspec_enabled)set_config_read_mode)fsspec_enabledrr ModuleTypec 8eZdZdZdededededdf fd ZxZS) ReduceLROnPlateauzCustom ReduceLROnPlateau scheduler that extends PyTorch's ReduceLROnPlateau. This class adds a `monitor` attribute to the standard PyTorch ReduceLROnPlateau to specify which metric should be tracked for learning rate adjustment. optimizermonitorargskwargsreturnNc:t||g|i|||_yN)super__init__r))selfr(r)r*r+ __class__s j/home/mcse/projects/srt_converter/srt-converter-venv/lib/python3.12/site-packages/pytorch_lightning/cli.pyr0zReduceLROnPlateau.__init__Ls! 4T4V4 ) __name__ __module__ __qualname____doc__rstrrr0 __classcell__r2s@r3r'r'Ds7)c#QTY]r4r'c $eZdZdZdddddededed ed ed d f fd Z ddee dee e e e ffee ee ee ee fdededed eef dZefddfdeeeeeedffdeded d fdZeddfdeeeedffdeded d fdZxZS)LightningArgumentParserzAExtension of jsonargparse's ArgumentParser for pytorch-lightning.z#Lightning Trainer command line toolPLF description env_prefix default_envr*r@rArBr+r,Nctsttt| ||||d|g|_i|_i|_y)aInitialize argument parser that supports configuration file input. For full details of accepted arguments see `ArgumentParser.__init__ `_. Args: description: Description of the tool shown when running ``--help``. env_prefix: Prefix for environment variables. Set ``default_env=True`` to enable env parsing. default_env: Whether to parse environment variables. r?N)"_JSONARGPARSE_SIGNATURES_AVAILABLEModuleNotFoundErrorr/r0 callback_keys _optimizers_lr_schedulers)r1r@rArBr*r+r2s r3r0z LightningArgumentParser.__init__`sI&2%)K(LN N $KJ\grkqr(*QSTVr4lightning_class. nested_key subclass_moderequiredc t|rt|ts t|}t|trt |t t ttfrgt |tr|jj||r|j||d|S|j||dt |t  dStd|d)aAdds arguments from a lightning class to a nested key of the parser. Args: lightning_class: A callable or any subclass of {Trainer, LightningModule, LightningDataModule, Callback}. nested_key: Name of the nested namespace to store arguments. subclass_mode: Whether allow any subclass of the given class. required: Whether the argument group is required. Returns: A list with the names of the class arguments added. F) fail_untypedrLT)rN instantiate sub_configszCannot add arguments from: zt. You should provide either a callable or a subclass of: Trainer, LightningModule, LightningDataModule, or Callback.)callable isinstancetyper issubclassrrrrrFappendadd_subclass_argumentsadd_class_argumentsr)r1rIrJrKrLs r3add_lightning_class_argsz0LightningArgumentParser.add_lightning_class_args{s2 O $Z-N1/BO ot , g8KXV2 /84""))*522?J]bmu2vv++" *?G DD , ()/):;J J  r4r( AUTOMATICoptimizer_classlink_toct|trtd|DsJt|tsJdddhd}t|tr|j ||fi|n|j ||fddi|||f|j|<y)adAdds arguments from an optimizer class to a nested key of the parser. Args: optimizer_class: Any subclass of :class:`torch.optim.Optimizer`. Use tuple to allow subclasses. nested_key: Name of the nested namespace to store arguments. link_to: Dot notation of a parser key to set arguments or AUTOMATIC. c3<K|]}t|tywr.)rTr.0os r3 z=LightningArgumentParser.add_optimizer_args..sIAz!Y/IFparamsrOrNskiprPTN)rRtupleallrTrrVrWrG)r1rZrJr[r+s r3add_optimizer_argsz*LightningArgumentParser.add_optimizer_argss ou -III IIoy9 9916X`Wa!b ou - 'D ' ' Nv N $D $ $_j ]d ]V\ ](7'A$r4 lr_schedulerlr_scheduler_classct|trtd|DsJt|tsJdddhd}t|tr|j ||fi|n|j ||fddi|||f|j|<y)aAdds arguments from a learning rate scheduler class to a nested key of the parser. Args: lr_scheduler_class: Any subclass of ``torch.optim.lr_scheduler.{_LRScheduler, ReduceLROnPlateau}``. Use tuple to allow subclasses. nested_key: Name of the nested namespace to store arguments. link_to: Dot notation of a parser key to set arguments or AUTOMATIC. c3<K|]}t|tywr.)rTLRSchedulerTypeTupler^s r3raz@LightningArgumentParser.add_lr_scheduler_args..sWqz!%9:WrbFr(rdrPTN)rRrfrgrTrmrVrWrH)r1rjrJr[r+s r3add_lr_scheduler_argsz-LightningArgumentParser.add_lr_scheduler_argss (% 0WDVWW WW02FG GG16XcWd!e (% 0 'D ' '(:J Q& Q $D $ $%7 `QU `Y_ `+=w*GJ'r4)FT)r5r6r7r8rr9boolr0r r rrrrrSlistrXrrfrhrmLRSchedulerTypernr:r;s@r3r=r=]sK A! WWW W  W  W WJ$-  S%:Mx WXX Y M  ! $ % N   - - - -  c- bQZ|%" BtId9os6J0KKLBB B  B6Sg(" H!/5#9M3N"NOHH H  Hr4r=cxeZdZdZ ddededededededd fd Ze d e d e d edd fdZ d e d e d edd fdZ y )SaveConfigCallbackafSaves a LightningCLI config to the log_dir when training starts. Args: parser: The parser object used to parse the configuration. config: The parsed configuration that will be saved. config_filename: Filename for the config file. overwrite: Whether to overwrite an existing config file. multifile: When input is multiple config files, saved config preserves this structure. save_to_log_dir: Whether to save the config to the log_dir. Raises: RuntimeError: If the config file already exists in the directory to avoid overwriting a previous run parserconfigconfig_filename overwrite multifilesave_to_log_dirr,Nc||_||_||_||_||_||_d|_|std|ts tdyy)NF save_configz`save_to_log_dir=False` only makes sense when subclassing SaveConfigCallback to implement `save_config` and it is desired to disable the standard behavior of saving to log_dir.) rtrurvrwrxry already_savedrrs ValueError)r1rtrurvrwrxrys r3r0zSaveConfigCallback.__init__sb  .""."}]DJ\']i (^r4trainer pl_modulestagec|jry|jr|j}|Jtjj ||j }t|}|jsa|jr|j|nd}|jj|}|r%t|jjd|d|jrQ|j!|d|j"j%|j&|d|j|j(|jr|j+|||d|_|jj|j|_y)NFz expected a to NOT exist. Aborting to avoid overwriting results of a previous run. You can delete the previous config file, set `LightningCLI(save_config_callback=None)` to disable config saving, or set `LightningCLI(save_config_kwargs={"overwrite": True})` to overwrite the config file.T)exist_ok) skip_nonerwrx)r|rylog_dirospathjoinrvrrwis_global_zeroisfilestrategy broadcast RuntimeErrorr2r5makedirsrtsaverurxr{)r1r~rrr config_pathfs file_existss r3setupzSaveConfigCallback.setupsM       ooG& &&'',,w0D0DEK(B>>8?8N8Nbii 4TY %..88E &>>223:k]Kww%% Gd 3   KKcgcqcq!  ! !   Wi 7!%D %--778J8JKr4cy)a Implement to save the config in some other place additional to the standard log_dir. Example: def save_config(self, trainer, pl_module, stage): if isinstance(trainer.logger, Logger): config = self.parser.dump(self.config, skip_none=False) # Required for proper reproducibility trainer.logger.log_hyperparams({"config": config}) Note: This method is only called on rank zero. This allows to implement a custom save config without having to worry about ranks or race conditions. Since it only runs on rank zero, any collective call will make the process hang waiting for a broadcast. If you need to make collective calls, implement the setup method instead. N)r1r~rrs r3r{zSaveConfigCallback.save_config-r4)z config.yamlFFT)r5r6r7r8r=rr9ror0rrrrr{rr4r3rsrss & - $'      .%LW%L%L%LQU%L%LN 7  s W[ r4rsc!~eZdZdZddededddeddddddfdeee e e de ffdeee e e de ffdee ed ee eefd ee ee deffd ee eefd eeefd eee eefe ee eefffde edededededededdf dZd e eefdee eefe eefffdZdedefdZdede eefde eefddfdZdeddfdZdeddfd Zdeddfd!Zdeddfd"Zede eeeffd#Zdededdfd$Zd%e d&ededefd'Z ededdfd(Z!dededdfd)Z"d@d*Z#d@d+Z$d@d,Z%d@d-Z&d@d.Z'd@d/Z(dedefd0Z)d1e eefd2e*e+defd3Z,d&eedefd4Z-e dAd5e d6e.d7ee/defd8Z0d&eeddfd9Z1dAd1e2d:ed;eedefd<Z3d&eddfd=Z4d&ede eeffd>Z5d@d?Z6y)B LightningCLIzIImplementation of a configurable command line tool for pytorch-lightning.NTF model_class.datamodule_classsave_config_callbacksave_config_kwargs trainer_classtrainer_defaultsseed_everything_default parser_kwargs parser_classsubclass_mode_modelsubclass_mode_datar*runauto_configure_optimizersload_from_checkpoint_supportr,c||_|xsi|_||_|xsi|_||_|xsi|_| |_||_||_|xst|_ |duxs| |_ ||_ |xst|_|duxs| |_|j!|j \}}|j#| |||j%|j&| |j)| r|j*dnd|_|j/|r|j1|j3|j5|j7|j,|j9|j,yy)aReceives as input pytorch-lightning classes (or callables which return pytorch-lightning classes), which are called / instantiated using a parsed configuration file and / or command line args. Parsing of configuration from environment variables can be enabled by setting ``parser_kwargs={"default_env": True}``. A full configuration yaml would be parsed from ``PL_CONFIG`` if set. Individual settings are so parsed from variables named for example ``PL_TRAINER__MAX_EPOCHS``. For more info, read :ref:`the CLI docs `. Args: model_class: An optional :class:`~pytorch_lightning.core.LightningModule` class to train on or a callable which returns a :class:`~pytorch_lightning.core.LightningModule` instance when called. If ``None``, you can pass a registered model with ``--model=MyModel``. datamodule_class: An optional :class:`~pytorch_lightning.core.datamodule.LightningDataModule` class or a callable which returns a :class:`~pytorch_lightning.core.datamodule.LightningDataModule` instance when called. If ``None``, you can pass a registered datamodule with ``--data=MyDataModule``. save_config_callback: A callback class to save the config. save_config_kwargs: Parameters that will be used to instantiate the save_config_callback. trainer_class: An optional subclass of the :class:`~pytorch_lightning.trainer.trainer.Trainer` class or a callable which returns a :class:`~pytorch_lightning.trainer.trainer.Trainer` instance when called. trainer_defaults: Set to override Trainer defaults or add persistent callbacks. The callbacks added through this argument will not be configurable from a configuration file and will always be present for this particular CLI. Alternatively, configurable callbacks can be added as explained in :ref:`the CLI docs `. seed_everything_default: Number for the :func:`~lightning_fabric.utilities.seed.seed_everything` seed value. Set to True to automatically choose a seed value. Setting it to False will avoid calling ``seed_everything``. parser_kwargs: Additional arguments to instantiate each ``LightningArgumentParser``. subclass_mode_model: Whether model can be any `subclass `_ of the given class. subclass_mode_data: Whether datamodule can be any `subclass `_ of the given class. args: Arguments to parse. If ``None`` the arguments are taken from ``sys.argv``. Command line style arguments can be given in a ``list``. Alternatively, structured config options can be given in a ``dict`` or ``jsonargparse.Namespace``. run: Whether subcommands should be added to run a :class:`~pytorch_lightning.trainer.trainer.Trainer` method. If set to ``False``, the trainer and model classes will be instantiated only. auto_configure_optimizers: Whether to automatically add default optimizer and lr_scheduler arguments. load_from_checkpoint_support: Whether ``save_hyperparameters`` should save the original parsed hyperparameters (instead of what ``__init__`` receives), such that it is possible for ``load_from_checkpoint`` to correctly instantiate classes even when using complex nesting and dependency injection. N subcommand)rrrrrrrrrr _model_classrrr_datamodule_classr_setup_parser_kwargs setup_parserparse_argumentsrt_parse_ckpt_pathrur _set_seed_add_instantiatorsbefore_instantiate_classesinstantiate_classesafter_instantiate_classes_run_subcommand)r1rrrrrrrrrrrr*rrr main_kwargssubparser_kwargss r3r0zLightningCLI.__init__Bsf@%9!"4":* 0 6B'>$*0b()B&&':?$/4$7#O:O:O:U:U:Wl$!Q[\`k[kJNA-ll,-''(9(97RVRjRj'k  ,  + +D,B,BFZ^ZqZq + r  + +&&d>U>U`e , ms C+C+c*|j||j||j||jrC|js|j t f|js|jt|j|yr.) rradd_arguments_to_parserrrGrhrrHrnrm!link_optimizers_and_lr_schedulersrs r3rzLightningCLI._add_argumentsst ,,V4 ))&1 $$V,  ) )%%))9,7((,,-AB ..v6r4cy)zImplement to add extra arguments to the parser or link arguments. Args: parser: The parser object to which arguments can be added Nrrs r3rz$LightningCLI.add_arguments_to_parserrr4c hdhdhdhddS)zDDefines the list of available subcommands and the arguments to skip.>r datamoduleval_dataloaderstrain_dataloaders>rr dataloaders)fitvalidatetestpredictrrr4r3rzLightningCLI.subcommandssS>:=   r4c i|_|j}t|jtr |jnt |j}|j D]t}t||}t|}|j|i}|jd||j||fi|} | |j|<|j|| |vy)z%Adds subcommands to the input parser.r@)rN) _subcommand_parsersrrRrrSrrgetattr_get_short_descriptiongetr_prepare_subcommand_parseradd_subcommand) r1rtr+parser_subcommandsrrfnr@rsubcommand_parsers r3rzLightningCLI._add_subcommandssGI #335#-T-?-?"FD  L_`d`r`rLs **, _J 3B04K%zz*b9   ' ' { C ? ? ? z n]m n 3DD $ $Z 0  - -j:KR] - ^ _r4klassrc |jdi|}|j|t|j|}|j |||}||j |<|S)N)rer)rrsetradd_method_argumentsr)r1rrr+rtreaddeds r3rz'LightningCLI._prepare_subcommand_parsersi!!!+F+ F#%()9)9);J)G%H++E:D+I8=))*5 r4ci|j|j}|jD]P\}\}}|dk(rt|tr|j ||2t |}|j |||Ry)z`Creates argument links for optimizers and learning rate schedulers that specified a ``link_to``.rY) compute_fnN)rGrHrrRrflink_arguments_add_class_path_generator)rtoptimizers_and_lr_schedulerskey class_typer[add_class_paths r3rz.LightningCLI.link_optimizers_and_lr_schedulerss(W&*<*<'V@U@U'V$*F*L*L*N O &C&*g+%*e,%%c73!::!F%%c7~%N Or4c |?ttjdkDr#tdtjddd|dt |t t fr|j||_y|j||_y)z?Parses command line arguments and stores it in ``self.config``.NzLightningCLI's args parameter is intended to run from within Python like if it were from the command line. To prevent mistakes it is not recommended to provide both args and command line arguments, got: sys.argv[1:]=z, args=.) lensysargvrrRdictr parse_objectru parse_args)r1rtr*s r3rzLightningCLI.parse_arguments&su  CHH  1  # ~WTF!=  dT9- . --d3DK ++D1DKr4cn|jjdsy|j|jjjd}|rt|j rt j |dd}|jdi}|jdd|syd |vr|jd |d }|jjd |ii} |jj||j|_yyy#t$r!tjjd wxYw) zcIf a checkpoint path is given, parse the hyperparameters from the checkpoint and update the config.rN ckpt_pathTcpu) weights_only map_locationhyper_parameters _instantiator _class_path class_path dict_kwargsrz-Parsing of ckpt_path hyperparameters failed! )rurrris_filetorchloadpoprtr SystemExitrstderrwrite)r1rckpthparamss r3rzLightningCLI._parse_ckpt_path3s {{|, KK 6 67;;KH i002::idODhh126G KK .'")++m"<#*{{--/ABG "kk66w L 39    !QR s +D *D4ct|drytj|jj |j ddd|_d|j vr)|j |j j|_yy)N config_dumpFyaml)skip_link_targetsrformatr)hasattrr safe_loadrtdumprurrr1s r3 _dump_configzLightningCLI._dump_configJsp 4 ' >> KK  T[[EU[a  b  4;; &#// 0F0FGD  'r4c|jjt|dt|j|j |jjt|dt|j |jy)Nrclir) subclassesr)rtadd_instantiator_InstantiatorFn_get_module_typerrrrrs r3rzLightningCLI._add_instantiatorsSst $$ ' 2 T.. /// % $$ & 1 T33 4.. % r4cy)z.get_automatic..s1`PQ*Q 2K1`s)rrRrfanyrU)rrA automaticr base_classr[s` r3 get_automaticzMLightningCLI._add_configure_optimizers_method_to_model..get_automaticsgI.6nn.> ***j'!*e4",Jk)c1`U_1`.`$$S)  *  r4rr`z{.add_configure_optimizers_method_to_model` expects at most one optimizer and one lr_scheduler to be 'AUTOMATIC', but found z. In this case the user is expected to link the argument groups and implement `configure_optimizers`, see https://lightning.ai/docs/pytorch/stable/common/lightning_cli.html#optimizers-and-learning-rate-schedulersr?z..configure_optimizers` will be overridden by `z.configure_optimizers`.r>)rr0r rSrfrr9rprrGrmrHrrr2r5r+r*rR_global_add_class_pathrrrinstantiate_class parametersrr?rr) r1rrtrG optimizers lr_schedulersrZoptimizer_initlr_scheduler_initrjr(rirs r3r,z6LightningCLI._add_configure_optimizers_method_to_modelsl-- j) dE$)$445 AEc5QVW[]bcgilcl]mWmQnpsQsKtFtAu  #Y #9f.@.@A %&:F4#3#3Z]C/513O^TN   !'!6!6}Q7G!H!K  $ $*:*:M!? %   $ )$7K)+tDK ??>IDKK (): ;-8DKK) *r4)r,Nr.)7r5r6r7r8rsrr=r r rSrr rrr9rrorArgsTyper0rfrrrrrrr staticmethodrrrrrrrr rrrrr-rprr1r0rLRSchedulerTypeUnionr?r,rr+rrTrrr4r3rr?sSS_ckoCU7;FM5948TX6M$)#(*.-1!c2eD$98CDX;Y$YZ[c2#5.A)BHSReMeDf)f#ghc2't,>'?@ c2 %T#s(^4 c2 T']HS'\,BBC c2#4S>2c2"'tSy!1c2 d38nd3S#X;N6O&O PQc223c2"c2!c2c2c2$(c2 '+!c2" #c2J-$sCx.-U4PSUXPX>[_`ceh`h[iKiEj- C,C -# -26sCx. -TXY\^aYaTb -  -  6M  RV  3Jt 7%< 7 7 .E $  c3s8m,  _'>_#_RV_&#QTYp O2I Od O O 2&= 2X 2RV 2.H  K2J JC JG J,4S>,d8n,Y`,24(3-44K4 pt+)+6?+OWXlOm+ ++(<EHSM<EVZ<E|J9J3J#JRUJ # $ ST#s(^ 9r4rrr,c:|jdz|jzS)Nr)r6r5)rs r3_class_path_from_classr`s  3 &)<)< <.add_class_path$s%j)<sp J$**9+<+<= *** (33 4J6ykB4HIIJs*:A.A))A.r8czt|r/t|tstj|j S|Sr.)rQrRrSinspect signaturereturn_annotation)r8s r3r'r'Hs.z%6  '999 Lr4rrcr|jd}|ddD]}||vs Jd|d||}|||d<y)Nrz Expected key z to be in data)split)rrr8rrs r3_set_dict_nestedr~NsV 99S>D #2YDy=M#n==yAwDbNr4c BeZdZdededdfdZdeedede d edef d Z y) r&r#rr,Nc ||_||_yr.r")r1r#rs r3r0z_InstantiatorFn.__init__Wsr4rr*applied_instantiation_linksr+cF|jj|jjj|ji}d|vr+d|di|jdi|jdi}|j D]k\}}|j |jds%|t|jdd}|j dr|tdd}t|||mt|d5||i|cdddS#1swYyxYw) Nr r rar rz init_args.z(pytorch_lightning.cli.instantiate_module)r instantiator) r#r rrrr startswithrr~r)r1rrr*r+rrr8s r3__call__z_InstantiatorFn.__call__[s( ((&&**488R8 7 " w|4++k2.++mR0G 6;;= 2JC>>TXXJa.1cTXXJa.)+,C~~l+#l+-. Wc5 1  2,C  /t.v.  / / /s DD ) r5r6r7rr9r0rSr%rrrrr4r3r&r&VsVLst/$//&* /  /  /r4r&ruc@td}d|vrE|j|dd|d|jDcic]\}}|dk7s ||c}}d}n|j|dd|j d|i}|j |}|j Scc}}w)NF) exit_on_errorr rm)rNr )rrVrrWrrrm)rrurtrrcfgrfs r3instantiate_moduler}s % 0F%%j(%O /-3\\^RTQqM?QAqDR  "":xe"L   x0 1C  % %c *D ;; Ss B Br.)Wrxrrcollections.abcr functoolsrrpathlibrtypesrtypingrr r r r rr lightning_utilities.core.importsr "lightning_utilities.core.rank_zeror torch.optimrtorch.optim.lr_schedulerrtyping_extensionsrpytorch_lightningr#lightning_fabric.utilities.cloud_iorrrrrr+pytorch_lightning.core.mixins.hparams_mixinr&pytorch_lightning.utilities.exceptionsr)pytorch_lightning.utilities.model_helpersr%pytorch_lightning.utilities.rank_zerorrDrq jsonargparserrrrr r! ImportErrorr#objectlocalsr%optimrir'rmr^rSrqrpr9rr\OptimizerCallableLRSchedulerCallabler=rsrr`rIrrfrJrr'r~r&rrr4r3rs $-:: =4!0&>ffVLC@%56X%Y"%'u-25TB "(FH "FH[ \ "   00BB $%67[*;;<[)40A+BBC E$s)T#s(^Y>? @hZ23 {E+?P2P,QQR~Hn~HB^ ^ BU9U9p=t== OS\\!)% 4S>0I*J!K\ #s(^\$8YKcSVh