&`i4DddlZddlZddlZddlZddlmZmZmZmZm Z m Z ddl m Z ddl mZmZmZmZddlmZmZmZmZmZmZddlmZddlmZmZ ddlZddl m!Z!dd l"m#Z#ej$j%Z%n#e&$r dZdZ!dZ#dZ%YnwxYwej'e(Z)Gd d eZ*dS) N)DictListOptionalSequenceTypeUnion)DEFAULT_METRIC)UNDEFINED_METRIC_MODEUNDEFINED_SEARCH_SPACEUNRESOLVED_SEARCH_SPACESearcher) CategoricalDomainFloatInteger LogUniform Quantized)parse_spec_vars) flatten_dictunflatten_dict) Optimizer)ConfiguredOptimizerceZdZdZ ddeeeeeefdee deee e fdee dee dee e f fd Z d Zdee dee d e d efd Zde d ee fdZ dde dee defdZde de fdZde fdZde fdZede d e fdZxZS)NevergradSearcha Uses Nevergrad to optimize hyperparameters. Nevergrad is an open source tool from Facebook for derivative free optimization. More info can be found at: https://github.com/facebookresearch/nevergrad. You will need to install Nevergrad via the following command: .. code-block:: bash $ pip install nevergrad Parameters: optimizer: Optimizer class provided from Nevergrad. See here for available optimizers: https://facebookresearch.github.io/nevergrad/optimizers_ref.html#optimizers This can also be an instance of a `ConfiguredOptimizer`. See the section on configured optimizers in the above link. optimizer_kwargs: Kwargs passed in when instantiating the `optimizer` space: Nevergrad parametrization to be passed to optimizer on instantiation, or list of parameter names if you passed an optimizer object. metric: The training result objective value attribute. If None but a mode was passed, the anonymous metric `_metric` will be used per default. mode: One of {min, max}. Determines whether objective is minimizing or maximizing the metric attribute. points_to_evaluate: Initial parameter suggestions to be run first. This is for when you already have some good parameters you want to run first to help the algorithm make better suggestions for future parameters. Needs to be a list of dicts containing the configurations. Tune automatically converts search spaces to Nevergrad's format: .. code-block:: python import nevergrad as ng config = { "width": tune.uniform(0, 20), "height": tune.uniform(-100, 100), "activation": tune.choice(["relu", "tanh"]) } current_best_params = [{ "width": 10, "height": 0, "activation": relu", }] ng_search = NevergradSearch( optimizer=ng.optimizers.OnePlusOne, metric="mean_loss", mode="min", points_to_evaluate=current_best_params) run(my_trainable, config=config, search_alg=ng_search) If you would like to pass the search space manually, the code would look like this: .. code-block:: python import nevergrad as ng space = ng.p.Dict( width=ng.p.Scalar(lower=0, upper=20), height=ng.p.Scalar(lower=-100, upper=100), activation=ng.p.Choice(choices=["relu", "tanh"]) ) ng_search = NevergradSearch( optimizer=ng.optimizers.OnePlusOne, space=space, metric="mean_loss", mode="min") run(my_trainable, search_alg=ng_search) N optimizeroptimizer_kwargsspacemetricmodepoints_to_evaluatecXt Jd|r|dvs Jdtt|||d|_d|_d|_|pi|_|d|_nIt|ts tdt|dt||_t|tri|rgt|\}}} |s| rPt t%jdt|||}t|t*rK|$t|tstd |jrtd ||_||_nct/j|rt3|t*st|t4r||_d|_||_ntd i|_|js|jr|dSdS) NzyNevergrad must be installed! You can install Nevergrad with the command: `pip install nevergrad`.)minmaxz`mode` must be 'min' or 'max'.)rrz5Invalid object type passed for `points_to_evaluate`: z6. Please pass a list of points (dictionaries) instead.r)parclszxIf you pass a configured optimizer to Nevergrad, either pass a list of parameter names or None as the `space` parameter.zmIf you pass in optimizer kwargs, either pass an `Optimizer` subclass or an instance of `ConfiguredOptimizer`.zlThe `optimizer` argument passed to NevergradSearch must be either an `Optimizer` or a `ConfiguredOptimizer`.)ngsuperr__init___space _opt_factory_nevergrad_opt_optimizer_kwargs_points_to_evaluate isinstancer ValueErrortypelistdictrloggerwarningr formatconvert_search_spacer _parametersinspectisclass issubclassr_live_trial_mapping_setup_nevergrad) selfrrrrrr resolved_vars domain_vars grid_vars __class__s ~/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/ray/tune/search/nevergrad/nevergrad_search.pyr(zNevergradSearch.__init__{s NN ( NN  L>)))+K))) ot$$--V$-GGG  "!1!7R  %'+D $ $.99 @G*++GGG  (,,>'?'?D $ eT " " 9u 94CE4J4J 1M;  9i 9+2wDJJOOO11%88 i + +  E4)@)@  ! %  - %D "+D   OI & & +5i+K+K   #6 7 7 !*D #D DKKD  $&   $$+ $  ! ! # # # # # $ $c2|jr|j|jfi|j|_|jdkrd|_n|jdkrd|_|j|jr t|_t|jdr|jj j r7|jj j rtd|j tdnG|j tdt|jj j d krtd|j 1|jjt|j krtd |jrNt#t|jd z d d D])}|j|j|(dSdS) Nr#gr"g?instrumentationz.Instrumented optimizers should use kwargs onlyz>Instrumented optimizers should provide None as parameter_nameszANon-instrumented optimizers should have a list of parameter_nameszTlen(parameters_names) must match optimizer dimension for non-instrumented optimizers)r*r)r,r+_mode _metric_op_metricr hasattrrEkwargsargsr/r7len dimensionr-rangesuggest)r=is rBr<z NevergradSearch._setup_nevergrads   "3$"3 ###5##D  :  "DOO Z5 !DO < DJ )DL 4&(9 : : W"29 W&6;W$%UVVV#/$20 #+$4t*:?@@AEE$%UVVV   'D,?,IS  N N - - <   # I3t7881Nevergrad does not support quantization. Dropped quantization.)lowerupperexponent)r}r~rF)choiceszMNevergrad does not support parameters of type `{}` with samplers of type `{}`) get_samplerr.rr3r4rrr&pLogr}r~matheScalarrset_integer_castingrChoice categoriesr/r5r0r[sampler)r{rs rB resolve_valuez;NevergradSearch.convert_search_space..resolve_valueNs((**G'9-- 0T"--//&%(( >gz22488$l&,$t{{V\{JJJFG,, >gz22,488$l$lQ.!%$*)++ , t{{ , ,*#&%''( FK00 >t{{6+<{===228&LL)4+?+?+H33 rCcTi|]$\}}d||%S)/)join).0pathr{rs rB z8NevergradSearch.convert_search_space..ts3WWW<4$v!6!6WWWrC)rr/rr Parameterr&rr)rVr>r?r@rrs @rBr6z$NevergradSearch.convert_search_space@s0?0E0E- {I  /  DD9990?0E0E- {I# &# Y# # # # LXWWW;WWWty!!5!!!rC)NNNNNN)NF)r[ __module__ __qualname____doc__rrrrrrrstrrr(r<boolrWrQrercrrrx staticmethodr6 __classcell__)rAs@rBrr(s$PPl +/26 $"37L$L$ )T)_.AA B L$ #4. L$ dIo./ L$ L$smL$%T$Z0L$L$L$L$L$L$\-I-I-I^sm+3C=BF  ;;;;;;:KP / / /%-d^ /CG / / / /XXTXXXX1C1111 *s**** 5"45"I5"5"5"\5"5"5"5"5"rCr)+r8loggingrrntypingrrrrrrray.tune.resultr ray.tune.searchr r r r ray.tune.search.samplerrrrrr!ray.tune.search.variant_generatorrray.tune.utils.utilrr nevergradr&nevergrad.optimizationrnevergrad.optimization.baserrr ImportError getLoggerr[r3rrrCrBrs >>>>>>>>>>>>>>>>****** >=====<<<<<<<< 000000??????II BIIII   8 $ $N"N"N"N"N"hN"N"N"N"N"sA// A?>A?