From a291abfab9c96e74ae4f7ecbb8b336d7a8c6f426 Mon Sep 17 00:00:00 2001 From: Qingyun Wu Date: Mon, 5 Jul 2021 21:17:26 -0400 Subject: [PATCH] Cha cha (#127) * unordered categorical * allow cost attribute to be None * tensorboardX version * quote * cfo cat * trunc * Update version.py * incumbent is normalized * python 3.9 * remove ConcurrencyLimiter * seed * estimator * update autovw notebook Co-authored-by: Chi Wang Co-authored-by: Qingyun Wu --- .github/workflows/python-package.yml | 5 +- README.md | 2 +- flaml/automl.py | 11 +--- flaml/model.py | 6 ++ flaml/searcher/blendsearch.py | 9 ++- flaml/searcher/cfo_cat.py | 31 +++++++++ flaml/searcher/flow2.py | 77 ++++++++++++++++------ flaml/searcher/search_thread.py | 10 +-- notebook/flaml_autovw.ipynb | 98 +++++++++++++++++++--------- setup.py | 2 +- 10 files changed, 177 insertions(+), 74 deletions(-) create mode 100644 flaml/searcher/cfo_cat.py diff --git a/.github/workflows/python-package.yml b/.github/workflows/python-package.yml index 7cef0a948..bf7e25d60 100644 --- a/.github/workflows/python-package.yml +++ b/.github/workflows/python-package.yml @@ -16,7 +16,7 @@ jobs: strategy: matrix: os: [ubuntu-latest, macos-latest, windows-2019] - python-version: [3.6, 3.7, 3.8] + python-version: [3.6, 3.7, 3.8, 3.9] steps: - uses: actions/checkout@v2 @@ -39,9 +39,10 @@ jobs: python -m pip install --upgrade pip pip install -e .[test] - name: If linux or mac, install ray - if: matrix.os == 'macOS-latest' || matrix.os == 'ubuntu-latest' + if: (matrix.os == 'macOS-latest' || matrix.os == 'ubuntu-latest') && matrix.python-version != '3.9' run: | pip install -e .[ray] + pip install 'tensorboardX<=2.2' - name: Lint with flake8 run: | # stop the build if there are Python syntax errors or undefined names diff --git a/README.md b/README.md index e5e84d48f..857f1816d 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ [![PyPI version](https://badge.fury.io/py/FLAML.svg)](https://badge.fury.io/py/FLAML) [![Build](https://github.com/microsoft/FLAML/actions/workflows/python-package.yml/badge.svg)](https://github.com/microsoft/FLAML/actions/workflows/python-package.yml) -![Python Version](https://img.shields.io/badge/3.6%20%7C%203.7%20%7C%203.8-blue) +![Python Version](https://img.shields.io/badge/3.6%20%7C%203.7%20%7C%203.8%20%7C%203.9-blue) [![Downloads](https://pepy.tech/badge/flaml/month)](https://pepy.tech/project/flaml) [![Join the chat at https://gitter.im/FLAMLer/community](https://badges.gitter.im/FLAMLer/community.svg)](https://gitter.im/FLAMLer/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge) diff --git a/flaml/automl.py b/flaml/automl.py index abf189d4c..359781091 100644 --- a/flaml/automl.py +++ b/flaml/automl.py @@ -981,15 +981,9 @@ class AutoML: self._retrained_config = {} est_retrain_time = next_trial_time = 0 best_config_sig = None - # use ConcurrencyLimiter to limit the amount of concurrency when - # using a search algorithm better = True # whether we find a better model in one trial if self._ensemble: self.best_model = {} - try: - from ray.tune.suggest import ConcurrencyLimiter - except ImportError: - from .searcher.suggestion import ConcurrencyLimiter if self._hpo_method in ('cfo', 'grid'): from flaml import CFO as SearchAlgo elif 'optuna' == self._hpo_method: @@ -1062,12 +1056,11 @@ class AutoML: metric='val_loss', mode='min', space=search_space, points_to_evaluate=points_to_evaluate, ) - search_state.search_alg = ConcurrencyLimiter(algo, - max_concurrent=1) + search_state.search_alg = algo else: search_space = None if self._hpo_method in ('bs', 'cfo'): - search_state.search_alg.searcher.set_search_properties( + search_state.search_alg.set_search_properties( config={ 'metric_target': self._state.best_loss, }, diff --git a/flaml/model.py b/flaml/model.py index 57fd34ca9..1363816cc 100644 --- a/flaml/model.py +++ b/flaml/model.py @@ -67,6 +67,12 @@ class BaseEstimator: ''' return self._model + @property + def estimator(self): + '''Trained model after fit() is called, or None before fit() is called + ''' + return self._model + def _preprocess(self, X): return X diff --git a/flaml/searcher/blendsearch.py b/flaml/searcher/blendsearch.py index 04d3b2bfb..fd9f35a3d 100644 --- a/flaml/searcher/blendsearch.py +++ b/flaml/searcher/blendsearch.py @@ -17,7 +17,7 @@ except ImportError: from .suggestion import OptunaSearch as GlobalSearch from .variant_generator import generate_variants from .search_thread import SearchThread -from .flow2 import FLOW2 as LocalSearch +from .flow2 import FLOW2 import logging logger = logging.getLogger(__name__) @@ -30,6 +30,7 @@ class BlendSearch(Searcher): cost_attr = "time_total_s" # cost attribute in result lagrange = '_lagrange' # suffix for lagrange-modified metric penalty = 1e+10 # penalty term for constraints + LocalSearch = FLOW2 def __init__(self, metric: Optional[str] = None, @@ -131,7 +132,7 @@ class BlendSearch(Searcher): self._gs = GlobalSearch(space=space, metric=metric, mode=mode) else: self._gs = None - self._ls = LocalSearch( + self._ls = self.LocalSearch( init_config, metric, mode, cat_hp_cost, space, prune_attr, min_resource, max_resource, reduction_factor, seed) self._init_search() @@ -277,7 +278,9 @@ class BlendSearch(Searcher): self._search_thread_pool[self._thread_count] = SearchThread( self._ls.mode, self._ls.create( - config, objective, cost=result[self.cost_attr]) + config, objective, + cost=result.get(self.cost_attr, 1)), + self.cost_attr ) thread_id = self._thread_count self._thread_count += 1 diff --git a/flaml/searcher/cfo_cat.py b/flaml/searcher/cfo_cat.py new file mode 100644 index 000000000..a6f884211 --- /dev/null +++ b/flaml/searcher/cfo_cat.py @@ -0,0 +1,31 @@ +'''! + * Copyright (c) 2021 Microsoft Corporation. All rights reserved. + * Licensed under the MIT License. See LICENSE file in the + * project root for license information. +''' +from .flow2 import FLOW2 +from .blendsearch import CFO + + +class FLOW2Cat(FLOW2): + '''Local search algorithm optimized for categorical variables + ''' + + def _init_search(self): + super()._init_search() + self.step_ub = 1 + self.step = self.STEPSIZE * self.step_ub + lb = self.step_lower_bound + if lb > self.step: + self.step = lb * 2 + # upper bound + if self.step > self.step_ub: + self.step = self.step_ub + self._trunc = self.dim + + +class CFOCat(CFO): + '''CFO optimized for categorical variables + ''' + + LocalSearch = FLOW2Cat diff --git a/flaml/searcher/flow2.py b/flaml/searcher/flow2.py index 60fce8d11..ad7a9d9cc 100644 --- a/flaml/searcher/flow2.py +++ b/flaml/searcher/flow2.py @@ -172,12 +172,12 @@ class FLOW2(Searcher): self._num_complete4incumbent = self._cost_complete4incumbent = 0 self._num_allowed4incumbent = 2 * self.dim self._proposed_by = {} # trial_id: int -> incumbent: Dict - self.step = self.STEPSIZE * np.sqrt(self.dim) + self.step_ub = np.sqrt(self.dim) + self.step = self.STEPSIZE * self.step_ub lb = self.step_lower_bound if lb > self.step: self.step = lb * 2 # upper bound - self.step_ub = np.sqrt(self.dim) if self.step > self.step_ub: self.step = self.step_ub # maximal # consecutive no improvements @@ -189,8 +189,11 @@ class FLOW2(Searcher): self._reset_times = 0 # record intermediate trial cost self._trial_cost = {} - self._same = False # whether the proposedd config is the same as best_config - self._init_phrase = True # initial phase to increase initial stepsize + self._same = False # whether the proposed config is the same as best_config + self._init_phase = True # initial phase to increase initial stepsize + self._trunc = 0 + # no truncation by default. when > 0, it means how many + # non-zero dimensions to keep in the random unit vector @property def step_lower_bound(self) -> float: @@ -215,7 +218,7 @@ class FLOW2(Searcher): if np.isinf(step_lb): step_lb = self.STEP_LOWER_BOUND else: - step_lb *= np.sqrt(self.dim) + step_lb *= self.step_ub return step_lb @property @@ -285,12 +288,14 @@ class FLOW2(Searcher): return unflatten_dict(config) def create(self, init_config: Dict, obj: float, cost: float) -> Searcher: - flow2 = FLOW2(init_config, self.metric, self.mode, self._cat_hp_cost, - unflatten_dict(self.space), self.prune_attr, - self.min_resource, self.max_resource, - self.resource_multiple_factor, self._seed + 1) + flow2 = self.__class__( + init_config, self.metric, self.mode, self._cat_hp_cost, + unflatten_dict(self.space), self.prune_attr, + self.min_resource, self.max_resource, + self.resource_multiple_factor, self._seed + 1) flow2.best_obj = obj * self.metric_op # minimize internally flow2.cost_incumbent = cost + self._seed += 1 return flow2 def normalize(self, config) -> Dict: @@ -315,10 +320,11 @@ class FLOW2(Searcher): elif key in self.incumbent: config_norm[key] = self.incumbent[ key] if value == self.best_config[ - key] else (self.incumbent[ - key] + 1) % self._unordered_cat_hp[key] + key] else ( + self.incumbent[key] + + 1.0 / self._unordered_cat_hp[key]) % 1 else: - config_norm[key] = 0 + config_norm[key] = 0.5 continue # Uniform/LogUniform/Normal/Base sampler = domain.get_sampler() @@ -365,7 +371,8 @@ class FLOW2(Searcher): config_denorm[key] = l[min(n - 1, int(np.floor(value * n)))] else: assert key in self.incumbent - if round(value) == self.incumbent[key]: + n = self._unordered_cat_hp[key] + if np.floor(value * n) == np.floor(self.incumbent[key] * n): config_denorm[key] = self.best_config[key] else: # ****random value each time!**** config_denorm[key] = self._random.choice( @@ -448,7 +455,11 @@ class FLOW2(Searcher): if self.step > self.step_ub: self.step = self.step_ub self._iter_best_config = self.trial_count_complete + if self._trunc: + self._trunc = min(self._trunc + 1, self.dim) return + elif self._trunc: + self._trunc = max(self._trunc >> 1, 1) proposed_by = self._proposed_by.get(trial_id) if proposed_by == self.incumbent: # proposed by current incumbent and no better @@ -494,8 +505,10 @@ class FLOW2(Searcher): # record the cost in case it is pruned and cost info is lost self._trial_cost[trial_id] = cost - def rand_vector_unit_sphere(self, dim) -> np.ndarray: + def rand_vector_unit_sphere(self, dim, trunc=0) -> np.ndarray: vec = self._random.normal(0, 1, dim) + if 0 < trunc < dim: + vec[np.abs(vec).argsort()[:dim - trunc]] = 0 mag = np.linalg.norm(vec) return vec / mag @@ -532,7 +545,7 @@ class FLOW2(Searcher): else: # propose a new direction self._direction_tried = self.rand_vector_unit_sphere( - self.dim) * self.step + self.dim, self._trunc) * self.step for i, key in enumerate(self._tunable_keys): move[key] += self._direction_tried[i] self._project(move) @@ -540,13 +553,14 @@ class FLOW2(Searcher): self._proposed_by[trial_id] = self.incumbent self._configs[trial_id] = (config, self.step) self._num_proposedby_incumbent += 1 - if self._init_phrase: + best_config = flatten_dict(self.best_config) + if self._init_phase: if self._direction_tried is None: if self._same: - # check if the new config is different from self.best_config + # check if the new config is different from best_config same = True for key, value in config.items(): - if key not in self.best_config or value != self.best_config[key]: + if key not in best_config or value != best_config[key]: same = False break if same: @@ -555,10 +569,10 @@ class FLOW2(Searcher): if self.step > self.step_ub: self.step = self.step_ub else: - # check if the new config is different from self.best_config + # check if the new config is different from best_config same = True for key, value in config.items(): - if key not in self.best_config or value != self.best_config[key]: + if key not in best_config or value != best_config[key]: same = False break self._same = same @@ -566,7 +580,7 @@ class FLOW2(Searcher): not self._resource or self._resource == self.max_resource): # check stuck condition if using max resource self._num_proposedby_incumbent -= 2 - self._init_phrase = False + self._init_phase = False if self.step >= self.step_lower_bound: # decrease step size self._oldK = self._K if self._K else self._iter_best_config @@ -574,6 +588,27 @@ class FLOW2(Searcher): self.step *= np.sqrt(self._oldK / self._K) else: return None + if self._init_phase: + return unflatten_dict(config) + if self._trunc == 1 and self._direction_tried is not None: + # random + for i, key in enumerate(self._tunable_keys): + if self._direction_tried[i] != 0: + for _, generated in generate_variants({'config': { + key: self.space[key] + }}): + if generated['config'][key] != best_config[key]: + config[key] = generated['config'][key] + return unflatten_dict(config) + break + else: + # check if config == best_config + if len(config) == len(best_config): + for key, value in best_config.items(): + if value != config[key]: + return unflatten_dict(config) + # print('move to', move) + self.incumbent = move return unflatten_dict(config) def _project(self, config): diff --git a/flaml/searcher/search_thread.py b/flaml/searcher/search_thread.py index 48f44394c..8640e6132 100644 --- a/flaml/searcher/search_thread.py +++ b/flaml/searcher/search_thread.py @@ -19,11 +19,11 @@ class SearchThread: '''Class of global or local search thread ''' - cost_attr = 'time_total_s' _eps = 1.0 def __init__(self, mode: str = "min", - search_alg: Optional[Searcher] = None): + search_alg: Optional[Searcher] = None, + cost_attr: Optional[str] = 'time_total_s'): ''' When search_alg is omitted, use local search FLOW2 ''' self._search_alg = search_alg @@ -40,6 +40,7 @@ class SearchThread: self.priority = self.speed = 0 self._init_config = True self.running = 0 # the number of running trials from the thread + self.cost_attr = cost_attr @classmethod def set_eps(cls, time_budget_s): @@ -108,9 +109,8 @@ class SearchThread: # under this thread self._init_config = False if result: - if self.cost_attr in result: - self.cost_last = result[self.cost_attr] - self.cost_total += self.cost_last + self.cost_last = result.get(self.cost_attr, 1) + self.cost_total += self.cost_last if self._search_alg.metric in result: obj = result[self._search_alg.metric] * self._metric_op if obj < self.obj_best1: diff --git a/notebook/flaml_autovw.ipynb b/notebook/flaml_autovw.ipynb index 508628e93..f9a1bb91d 100644 --- a/notebook/flaml_autovw.ipynb +++ b/notebook/flaml_autovw.ipynb @@ -64,7 +64,9 @@ { "output_type": "stream", "name": "stdout", - "text": "(36203, 17) (36203,)\n" + "text": [ + "(36203, 17) (36203,)\n" + ] } ], "source": [ @@ -88,7 +90,7 @@ }, { "cell_type": "code", - "execution_count": 2, + "execution_count": 3, "metadata": { "tags": [] }, @@ -96,7 +98,9 @@ { "output_type": "stream", "name": "stdout", - "text": "openml example: 8.170000076293945 [1.0000e+01 7.0000e+00 3.0000e+00 4.0000e+00 nan 6.3300e+00\n 1.3600e-01 7.3300e+00 7.0100e+00 6.9800e+00 3.0000e-03 7.0000e+00\n 9.7000e+00 1.2300e+01 1.0217e+03 0.0000e+00 5.8000e+01]\nvw example: 8.170000076293945 |a 0:10.000000 1:7.000000|b 2:3.000000 3:4.000000|c 4:nan 5:6.330000|d 6:0.136000 7:7.330000|e 8:7.010000 9:6.980000|f 10:0.003000 11:7.000000|g 12:9.700000 13:12.300000|h 14:1021.700012 15:0.000000|i 16:58.000000\n" + "text": [ + "openml example: 8.170000076293945 [1.0000e+01 7.0000e+00 3.0000e+00 4.0000e+00 nan 6.3300e+00\n 1.3600e-01 7.3300e+00 7.0100e+00 6.9800e+00 3.0000e-03 7.0000e+00\n 9.7000e+00 1.2300e+01 1.0217e+03 0.0000e+00 5.8000e+01]\nvw example: 8.170000076293945 |a 0:10.000000 1:7.000000|b 2:3.000000 3:4.000000|c 4:nan 5:6.330000|d 6:0.136000 7:7.330000|e 8:7.010000 9:6.980000|f 10:0.003000 11:7.000000|g 12:9.700000 13:12.300000|h 14:1021.700012 15:0.000000|i 16:58.000000\n" + ] } ], "source": [ @@ -139,7 +143,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 4, "metadata": {}, "outputs": [], "source": [ @@ -149,7 +153,6 @@ " \"\"\"\n", " print('Online learning for', iter_num, 'steps...')\n", " loss_list = []\n", - " y_predict_list = []\n", " for i in range(iter_num):\n", " vw_x = vw_examples[i]\n", " y_true = float(vw_examples[i].split('|')[0])\n", @@ -160,7 +163,6 @@ " # calculate one step loss\n", " loss = mean_squared_error([y_pred], [y_true])\n", " loss_list.append(loss)\n", - " y_predict_list.append([y_pred, y_true])\n", " return loss_list\n", "\n", "max_iter_num = 10000 # or len(vw_examples)" @@ -176,7 +178,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 5, "metadata": { "tags": [] }, @@ -184,13 +186,16 @@ { "output_type": "stream", "name": "stdout", - "text": "Online learning for 10000 steps...\nFinal progressive validation loss of vanilla vw: 15.180878192648041\n" + "text": [ + "Online learning for 10000 steps...\n", + "Final progressive validation loss of vanilla vw: 15.18087237487917\n" + ] } ], "source": [ "from vowpalwabbit import pyvw\n", "''' create a vanilla vw instance '''\n", - "vanilla_vw = pyvw.vw()\n", + "vanilla_vw = pyvw.vw('--quiet')\n", "\n", "# online learning with vanilla VW\n", "loss_list_vanilla = online_learning_loop(max_iter_num, vw_examples, vanilla_vw)\n", @@ -207,7 +212,7 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 6, "metadata": { "slideshow": { "slide_type": "slide" @@ -218,7 +223,16 @@ { "output_type": "stream", "name": "stderr", - "text": "Seed namespaces (singletons and interactions): ['e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nCreated challengers from champion |\nNew challenger size 37, ['|ah', '|ch', '|df', '|ef', '|ag', '|bg', '|be', '|eh', '|hi', '|cd', '|ci', '|eg', '|bh', '|ad', '|bi', '|ab', '|cg', '|bc', '|gi', '|ai', '|cf', '|ei', '|dg', '|ac', '|af', '|ce', '|ae', '|de', '|fi', '|bd', '|gh', '|bf', '|dh', '|di', '|fh', '|fg', '|']\nOnline learning for 10000 steps...\nSeed namespaces (singletons and interactions): ['dh', 'e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nCreated challengers from champion |dh\nNew challenger size 43, ['|dh_ei', '|bd_dh', '|cdh_dh', '|ac_dh', '|bh_dh', '|ab_dh', '|dh_gi', '|cg_dh', '|bf_dh', '|dh_dhi', '|deh_dh', '|dh_fi', '|ad_dh', '|dh_hi', '|dh_eg', '|bdh_dh', '|dh_eh', '|ag_dh', '|dh', '|de_dh', '|dgh_dh', '|bc_dh', '|cd_dh', '|dh_ef', '|cf_dh', '|dh_di', '|bi_dh', '|ah_dh', '|dh_fh', '|ce_dh', '|ae_dh', '|adh_dh', '|df_dh', '|ch_dh', '|dh_fg', '|ai_dh', '|ci_dh', '|dh_gh', '|dfh_dh', '|af_dh', '|dg_dh', '|be_dh', '|bg_dh']\nFinal progressive validation loss of autovw: 10.744201540966063\n" + "text": [ + "Seed namespaces (singletons and interactions): ['g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "Created challengers from champion ||\n", + "New challenger size 37, ['|ah|', '|eg|', '|gi|', '|ag|', '|de|', '|ei|', '|eh|', '|fg|', '|cf|', '|hi|', '|bf|', '|cd|', '|ai|', '|ef|', '|cg|', '|ch|', '|ad|', '|bc|', '|gh|', '|bh|', '|ci|', '|fh|', '|bg|', '|be|', '|bd|', '|fi|', '|bi|', '|df|', '|ac|', '|ae|', '|dg|', '|af|', '|di|', '|ce|', '|dh|', '|ab|', '||']\n", + "Online learning for 10000 steps...\n", + "Seed namespaces (singletons and interactions): ['ce', 'g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "Created challengers from champion |ce|\n", + "New challenger size 43, ['|be_ce|', '|bce_ce|', '|ce_ei|', '|ce_ceg|', '|ce_fh|', '|ce_gh|', '|ce_cef|', '|cd_ce|', '|ce_cg|', '|cde_ce|', '|ce_cf|', '|bd_ce|', '|ae_ce|', '|ce_gi|', '|ce_ci|', '|ab_ce|', '|ce_fg|', '|ce_di|', '|bi_ce|', '|ce_de|', '|ce_eg|', '|ce_dg|', '|ce_hi|', '|ai_ce|', '|ag_ce|', '|ac_ce|', '|bh_ce|', '|ce_ch|', '|ce|', '|ace_ce|', '|ah_ce|', '|af_ce|', '|bc_ce|', '|ce_dh|', '|ce_ef|', '|ad_ce|', '|ce_df|', '|ce_cei|', '|ce_eh|', '|bg_ce|', '|ce_ceh|', '|bf_ce|', '|ce_fi|']\n", + "Final progressive validation loss of autovw: 8.718817421944529\n" + ] } ], "source": [ @@ -226,7 +240,9 @@ "from flaml import AutoVW\n", "\n", "'''create an AutoVW instance for tuning namespace interactions'''\n", - "autovw_ni = AutoVW(max_live_model_num=5, search_space={'interactions': AutoVW.AUTOMATIC})\n", + "# configure both hyperparamters to tune, e.g., 'interactions', and fixed arguments about the online learner,\n", + "# e.g., 'quiet' in the search_space argument.\n", + "autovw_ni = AutoVW(max_live_model_num=5, search_space={'interactions': AutoVW.AUTOMATIC, 'quiet': ''})\n", "\n", "# online learning with AutoVW\n", "loss_list_autovw_ni = online_learning_loop(max_iter_num, vw_examples, autovw_ni)\n", @@ -242,15 +258,15 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 7, "metadata": {}, "outputs": [ { "output_type": "display_data", "data": { "text/plain": "
", - "image/svg+xml": "\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n", - "image/png": "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\n" + "image/svg+xml": "\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n", + "image/png": "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\n" }, "metadata": { "needs_background": "light" @@ -286,7 +302,7 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 8, "metadata": { "tags": [] }, @@ -294,14 +310,25 @@ { "output_type": "stream", "name": "stderr", - "text": "Seed namespaces (singletons and interactions): ['e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nNo low-cost init config given to the search algorithm.For cost-frugal search, consider providing init values for cost-related hps via 'init_config'.\nCreated challengers from champion ||0.5\nNew challenger size 39, ['|dh|0.5', '|ci|0.5', '|bd|0.5', '|bh|0.5', '|ei|0.5', '|ch|0.5', '|bg|0.5', '|bc|0.5', '|cd|0.5', '|ag|0.5', '|eh|0.5', '|hi|0.5', '|dg|0.5', '|fi|0.5', '|ad|0.5', '|cf|0.5', '|ce|0.5', '|be|0.5', '|ab|0.5', '|ah|0.5', '|fh|0.5', '|di|0.5', '|gi|0.5', '|bf|0.5', '|de|0.5', '|ac|0.5', '|ai|0.5', '|df|0.5', '|cg|0.5', '|ae|0.5', '|fg|0.5', '|ef|0.5', '|eg|0.5', '|gh|0.5', '|af|0.5', '|bi|0.5', '||0.05358867312681484', '||1.0', '||0.5']\nOnline learning for 10000 steps...\nSeed namespaces (singletons and interactions): ['e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nNo low-cost init config given to the search algorithm.For cost-frugal search, consider providing init values for cost-related hps via 'init_config'.\nCreated challengers from champion ||1.0\nNew challenger size 38, ['|bf|1.0', '|ab|1.0', '|fg|1.0', '|bg|1.0', '|ad|1.0', '|fi|1.0', '|be|1.0', '|gi|1.0', '|df|1.0', '|de|1.0', '|cg|1.0', '|hi|1.0', '|di|1.0', '|ei|1.0', '|ai|1.0', '|bc|1.0', '|af|1.0', '|ef|1.0', '|ag|1.0', '|dh|1.0', '|fh|1.0', '|cd|1.0', '|dg|1.0', '|gh|1.0', '|ah|1.0', '|eg|1.0', '|ci|1.0', '|ch|1.0', '|eh|1.0', '|ac|1.0', '|ce|1.0', '|bi|1.0', '|bd|1.0', '|ae|1.0', '|cf|1.0', '|bh|1.0', '||0.10717734625362937', '||0.3273795141019504']\nSeed namespaces (singletons and interactions): ['de', 'e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nNo low-cost init config given to the search algorithm.For cost-frugal search, consider providing init values for cost-related hps via 'init_config'.\nCreated challengers from champion |de|1.0\nNew challenger size 45, ['|cf_de|1.0', '|ci_de|1.0', '|cd_de|1.0', '|ac_de|1.0', '|de_dh|1.0', '|ab_de|1.0', '|de_ef|1.0', '|ce_de|1.0', '|de_hi|1.0', '|bg_de|1.0', '|de_fi|1.0', '|ah_de|1.0', '|de_dg|1.0', '|de_fg|1.0', '|ai_de|1.0', '|de_gh|1.0', '|bh_de|1.0', '|ch_de|1.0', '|de|1.0', '|af_de|1.0', '|de_deg|1.0', '|de_eh|1.0', '|de_eg|1.0', '|de_di|1.0', '|de_ei|1.0', '|ag_de|1.0', '|ae_de|1.0', '|de_deh|1.0', '|be_de|1.0', '|de_fh|1.0', '|cg_de|1.0', '|bf_de|1.0', '|bi_de|1.0', '|ad_de|1.0', '|ade_de|1.0', '|de_def|1.0', '|bde_de|1.0', '|cde_de|1.0', '|de_df|1.0', '|bc_de|1.0', '|de_dei|1.0', '|bd_de|1.0', '|de_gi|1.0', '|de|0.10717734625362937', '|de|0.3273795141019504']\nFinal progressive validation loss of autovw_nilr: 6.271218842008241\n" + "text": [ + "Seed namespaces (singletons and interactions): ['g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "No low-cost partial config given to the search algorithm. For cost-frugal search, consider providing low-cost values for cost-related hps via 'low_cost_partial_config'.\n", + "Created challengers from champion ||0.5|\n", + "New challenger size 39, ['|gi|0.5|', '|af|0.5|', '|df|0.5|', '|gh|0.5|', '|ae|0.5|', '|di|0.5|', '|be|0.5|', '|ac|0.5|', '|hi|0.5|', '|de|0.5|', '|ef|0.5|', '|bc|0.5|', '|cf|0.5|', '|dg|0.5|', '|fg|0.5|', '|bh|0.5|', '|ei|0.5|', '|ce|0.5|', '|bf|0.5|', '|ah|0.5|', '|ad|0.5|', '|bg|0.5|', '|bd|0.5|', '|ab|0.5|', '|bi|0.5|', '|eg|0.5|', '|ai|0.5|', '|eh|0.5|', '|dh|0.5|', '|cd|0.5|', '|fi|0.5|', '|ci|0.5|', '|ag|0.5|', '|fh|0.5|', '|ch|0.5|', '|cg|0.5|', '||0.05358867312681484|', '||1.0|', '||0.5|']\n", + "Online learning for 10000 steps...\n", + "Seed namespaces (singletons and interactions): ['g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "No low-cost partial config given to the search algorithm. For cost-frugal search, consider providing low-cost values for cost-related hps via 'low_cost_partial_config'.\n", + "Created challengers from champion ||1.0|\n", + "New challenger size 50, ['|gi|0.5|', '|af|0.5|', '|df|0.5|', '|gh|0.5|', '|ae|0.5|', '|di|0.5|', '|be|0.5|', '|ac|0.5|', '|hi|0.5|', '|de|0.5|', '|ef|0.5|', '|bc|0.5|', '|dh|1.0|', '|ah|1.0|', '|cd|1.0|', '|bh|1.0|', '|bi|1.0|', '|ab|1.0|', '|gi|1.0|', '|bg|1.0|', '|bd|1.0|', '|eh|1.0|', '|af|1.0|', '|hi|1.0|', '|cf|1.0|', '|ei|1.0|', '|ef|1.0|', '|ai|1.0|', '|ch|1.0|', '|gh|1.0|', '|fg|1.0|', '|ad|1.0|', '|ci|1.0|', '|bc|1.0|', '|ag|1.0|', '|df|1.0|', '|dg|1.0|', '|de|1.0|', '|di|1.0|', '|cg|1.0|', '|be|1.0|', '|eg|1.0|', '|ce|1.0|', '|fi|1.0|', '|ae|1.0|', '|bf|1.0|', '|fh|1.0|', '|ac|1.0|', '||0.10717734625362937|', '||0.3273795141019504|']\n", + "Final progressive validation loss of autovw_nilr: 7.611900319489723\n" + ] } ], "source": [ "from flaml.tune import loguniform\n", "''' create another AutoVW instance for tuning namespace interactions and learning rate'''\n", "# set up the search space and init config\n", - "search_space_nilr = {'interactions': AutoVW.AUTOMATIC, 'learning_rate': loguniform(lower=2e-10, upper=1.0)}\n", + "search_space_nilr = {'interactions': AutoVW.AUTOMATIC, 'learning_rate': loguniform(lower=2e-10, upper=1.0), 'quiet': ''}\n", "init_config_nilr = {'interactions': set(), 'learning_rate': 0.5}\n", "# create an AutoVW instance\n", "autovw_nilr = AutoVW(max_live_model_num=5, search_space=search_space_nilr, init_config=init_config_nilr)\n", @@ -321,7 +348,7 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 10, "metadata": { "tags": [] }, @@ -330,8 +357,8 @@ "output_type": "display_data", "data": { "text/plain": "
", - "image/svg+xml": "\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n", - "image/png": "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\n" + "image/svg+xml": "\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n", + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfQAAAFzCAYAAADIY/vqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8GearUAAAgAElEQVR4nOzdd3hc1Z3/8feZURn1LlmWZUvuveNOMSU2HQIBnA4hbLLLkpBkFzYbCMnubze7SViSDdmEkAQ2ydoQJxBq6MYGG3DBveIiW5at3stoNHN+f9yRbGPZHlsaj0b6vJ5HjzR37p35jnjwR+fcU4y1FhEREYlurkgXICIiIj2nQBcREekHFOgiIiL9gAJdRESkH1Cgi4iI9AMKdBERkX4gJtIF9ER2drYtKiqKdBkiIiLnxfr166ustTndPRfVgV5UVMS6desiXYaIiMh5YYwpOdVz6nIXERHpBxToIiIi/YACXUREpB+I6nvoIiISGp/PR2lpKW1tbZEuRULg8XgYMmQIsbGxIV+jQBcRGQBKS0tJSUmhqKgIY0yky5HTsNZSXV1NaWkpxcXFIV+nLncRkQGgra2NrKwshXkUMMaQlZV11r0pCnQRkQFCYR49zuW/lQJdRETCbuHChbzyyisnHHvkkUf46le/elav89xzz/GDH/wAgIceeogf/ehHAHzxi19k+fLlp7zuySefZMmSJSccq6qqIicnh7/85S/ccMMNXcf//d//nZEjR3Y9fv7557nuuuvOqs5IUKCLiEjYLVmyhGXLlp1wbNmyZSeF7Jlcd9113H///Wf9/jfeeCOvvfYaLS0tXceWL1/Otddey7x583jvvfe6jq9Zs4bU1FQqKioAWL16NfPmzTvr9zzfFOgiIhJ2N998My+++CLt7e0AHDhwgLKyMpYuXcrMmTOZMGEC3/3ud7vOLyoq4rvf/S7Tp09n0qRJ7Ny5E4AnnniCu++++7Tv9f3vf58LLriAiRMnctddd2GtJTU1lYsvvpjnn3++67zOPyhycnJITU3lo48+AuDw4cPcdNNNrF69GnACff78+b36+wgHjXIXERlgvvf8NraXNfTqa44fnMp3r51wyuczMzOZNWsWL7/8Mtdffz3Lli3jlltu4dvf/jaZmZn4/X4uu+wyNm/ezOTJkwHIzs5mw4YN/PznP+dHP/oRjz/+eEi13H333Tz44IMAfO5zn+OFF17g2muvZcmSJfzhD3/g1ltvpaysjN27d3PppZcCMH/+fFavXo3f72fUqFHMmTOHV155hWuuuYZNmzZxwQUX9PA3FH5qoQe9/v4fefq1n0a6DBGRfuv4bvfO1vHTTz/N9OnTmTZtGtu2bWP79u1d53/yk58EYMaMGRw4cCDk93nrrbeYPXs2kyZN4s0332Tbtm0AXH311bz77rs0NDTw9NNPc9NNN+F2uwGYN28eq1evZvXq1cydO5dZs2bx/vvv8+GHHzJ27Fg8Hk8v/RbCRy30oKc3PcJ+Vz23cE+kSxERCavTtaTD6frrr+fee+9lw4YNtLS0kJmZyY9+9CPWrl1LRkYGX/ziF0+YqhUfHw+A2+2mo6MjpPdoa2vjb//2b1m3bh2FhYU89NBDXa+ZkJDA4sWLeeaZZ1i2bBkPP/xw13Xz58/nv//7v/H7/Xz5y18mJSWFtrY2VqxYERX3z0Et9C4uDBYb6TJERPqt5ORkFi5cyB133MGSJUtoaGggKSmJtLQ0ysvLefnll3v8Hp3hnZ2dTVNT00kj35csWcLDDz9MeXk5c+fO7To+btw4ysrKeOedd5g2bRoAU6dO5Re/+EVU3D8HBfpxDAFN0RQRCaslS5awadMmlixZwpQpU5g2bRpjx47l05/+dK8EZ3p6Ol/+8peZOHEiixYtOune9xVXXEFZWRm33nrrCXO9jTHMnj2brKysruVW586dy759+6KmhW6sjd5W6cyZM21v7Yf+t49dxHZ3NSu+tK1XXk9EpC/ZsWMH48aNi3QZcha6+29mjFlvrZ3Z3flqoQcZXOpwFxGRqKVADzJGXe4iIhK9FOhBzqA4ERGR6KRADzK4CES6CBERkXOkQA9yGZe63EVEJGop0IMMRi10ERGJWgr0ICfQ1UQXEQmnZ599FmNM12Yrp/PII4+csDtad26//XZ++ctfnvQeV155Jffeey+PPPJI1/FFixZx5513dj3+5je/ecJqceBsyZqYmNi10xo4C+J093Nfo0APchlNWxMRCbelS5eyYMECli5desZzQwn0023L2rnhCkAgEKCqqqprXXc49bao2dnZ/PjHPw7l4/QpCvQgo5XiRETCqqmpiXfeeYdf//rXXSG8YsUKrrnmmq5z7r77bp544gl++tOfUlZWxsKFC1m4cCHg/DEwadIkJk6cyH333QfAZZddxs6dOzly5AgAzc3NvP7669xwww3MmzePNWvWALBt2zYmTpxISkoKtbW1eL1eduzYwfTp00+q84477uCpp56ipqYmrL+P3qbNWYJcxq176CIyMLx8Pxzd0ruvOWgSXPmD057yl7/8hcWLFzN69GiysrJYv379Kc+95557ePjhh3nrrbfIzs6mrKyM++67j/Xr15ORkcEnPvEJnn32WW644QZuuukmnn76ab72ta/x/PPPc8kll5CamkpqaioxMTEcPHiwaxe1w4cPs2bNGtLS0pg0aRJxcXE8+OCDzJw5k+uuuw5wutXvuOMOfvKTn/C9732vV39N4aQWepCzsIwh4PdHuhQRkX5p6dKl3HbbbQDcdtttIXW7d1q7di2XXHIJOTk5xMTE8JnPfIaVK1cC3W/L2unj26LOnTu363Hn2vHf//73u8K80z333MOTTz5JY2Njjz7z+aQWepAr+LdNwAZw4Y5wNSIiYXSGlnQ41NTU8Oabb7JlyxaMMfj9fowxXH/99QQCx/pHj98+NVTz5s3jyJEjbNq0idWrV59wT73zPvqWLVuYOHEihYWF/PjHPyY1NZXbb7/9lK+Znp7Opz/9aR599NGzridS1EIPMsFfRXtHe4QrERHpf5YvX87nPvc5SkpKOHDgAIcOHaK4uJhAIMD27dvxer3U1dXxxhtvdF2TkpLS1UKeNWsWb7/9NlVVVfj9fpYuXcrFF18MOD2st956K1/4whe48sor8Xg8Xa8xb948XnjhBTIzM3G73WRmZlJXV8eaNWvOuIvaN77xDX75y1+GvBd7pCnQg4xxfhV+vy/ClYiI9D9Lly7lxhtvPOHYTTfdxLJly7jllluYOHEit9xyS9de5AB33XUXixcvZuHCheTn5/ODH/yAhQsXMmXKFGbMmMH111/fde7x27Ieb9KkSVRVVTFnzpwTjqWlpZGdnQ3Agw8+yHPPPXdSzdnZ2dx44414vd5e+R2Em7ZPDfruk7fyZ7az6pNvkZ6S3SuvKSLSV2j71Oij7VPPkSu40b1PXe4iIhKFFOhBxjgD4fwa5S4iIlFIgR7UOcrd16F76CIiEn0U6EGdLfRAIDpGM4qIiBxPgR6ke+giIhLNFOhBndPWAlYtdBERiT4K9CCXBsWJiITdQNo+9ZJLLuHjU6tXrFhBWloaU6dOZezYsXzrW98K+fXORIEe5Aq20H1aWEZEJGz64/apTzzxBA899FDI51944YVs3LiRDz/8kBdeeIF333035GtPR4Ee5NKgOBGRsNL2qSdKSEhg6tSpHD58uFdeT5uzBHW20KNlzV4RkXP1Hx/8BztrztzlfTbGZo7lvln3nfYcbZ96otraWvbs2cNFF13UK6+nFnqQWugiIuHVn7ZPra6uZurUqUydOpUHH3yQX/ziF12Pt2w5/V7zq1atYsqUKRQUFLBo0SIGDRoU8u/hdNRCD+oc5d4R0D10EenfztSSDof+tn1qVlYWGzduBJx76AcOHAj5PvqFF17ICy+8wP79+5kzZw633HILU6dODf0Dn4Ja6EEuV2cLXaPcRUR6m7ZPPVlxcTH3338///Ef/9Err6dAD3IZp7PCr0AXEel1A3X71KuvvpohQ4YwZMgQPvWpT530/Fe+8hVWrlzJgQMHzvk9Omn71KBfPvttflb/PD8Z910unXVzr7ymiEhfoe1To4+2Tz1Hbpda6CIiEr0U6EFdK8VplLuIiEQhBXqQyxVcy10tdBERiUIK9KDOFrqmrYlIfxXNY6YGmnP5b6VAD+oMdKsWuoj0Qx6Ph+rqaoV6FLDWUl1dfcL0u1BoYZkgV3BQXMAGznCmiEj0GTJkCKWlpVRWVka6FAmBx+NhyJAhZ3WNAj2oay13vwbFiUj/ExsbS3FxcaTLkDBSl3uQ2xULQMAq0EVEJPoo0IM6R7lrHrqIiESjPtPlboy5AbgaSAV+ba199Xy+f0zwHrq1CnQREYk+YW2hG2N+Y4ypMMZs/djxxcaYXcaYj4wx9wNYa5+11n4Z+Apwazjr6rZWlxaWERGR6BXuLvcngMXHHzDGuIFHgSuB8cASY8z44075TvD588rduduaWugiIhKFwhro1tqVQM3HDs8CPrLW7rPWtgPLgOuN4z+Al621G071msaYu4wx64wx63pz+kXXoLiA5miKiEj0icSguALg0HGPS4PH/h64HLjZGPOVU11srX3MWjvTWjszJyen14rqnLamUe4iIhKN+sygOGvtT4GfRur9Y9yd09bU5S4iItEnEi30w0DhcY+HBI9FlKtrUJxWihMRkegTiUBfC4wyxhQbY+KA24DnIlDHCTpb6Jq2JiIi0Sjc09aWAmuAMcaYUmPMl6y1HcDdwCvADuBpa+22cNYRCpemrYmISBQL6z10a+2SUxx/CXgpnO99tjoD3WpzFhERiUJa+jUoxhUHaLc1ERGJTgr0IJdbC8uIiEj0UqAHHRsUpxa6iIhEHwV6kDu4OYtfLXQREYlCURnoxphrjTGP1dfX99prxsZ07ramFrqIiESfqAx0a+3z1tq70tLSeu01XcYJdN1DFxGRaBSVgR4ObndnoGtzFhERiT4K9KAYtzNtTV3uIiISjRToQe7gtLVV3lPu3CoiItJnKdCDYmOcaWtlsSbClYiIiJw9BXpQjCs20iWIiIicMwV6UEyMAl1ERKKXAj3I7Vagi4hI9FKgB8W4wrrxnIiISFgp0INi3Ap0ERGJXlEZ6OFY+rVztzUREZFoFJWBHo6lX48X8Gv5VxERiS5RGejh1uptiXQJIiIiZ0WB3o0Wb2OkSxARETkrCvRutLU1R7oEERGRs6JA70ZruwJdRESiiwK9G226hy4iIlFGgd4NBbqIiEQbBXo3vD4FuoiIRBcFejfa2lsjXYKIiMhZUaAfZ77XWaimvUOBLiIi0UWBfpyLht0IQLuvLcKViIiInB0F+nHiYjwAtHco0EVEJLpEZaCHY3MWgLhYJ9B9Hd5efV0REZFwi8pAD9fmLPGxCQC0K9BFRCTKRGWgh0tnoPv86nIXEZHookA/TnxcZ6C3R7gSERGRs3PGQDfG/KcxJtUYE2uMecMYU2mM+ez5KO58i49LAqDDry53ERGJLqG00D9hrW0ArgEOACOBfwhnUZFyrMtdLXQREYkuoQR6TPD71cAfrbW9O7S8D0kIdrn/b/vqCFciIiJydmLOfAovGGN2Aq3AV40xOUC/HDXm8SRHugQREZFzcsYWurX2fmAeMNNa6wOagevDXVgkJHqSIl2CiIjIOQllUNynAJ+11m+M+Q7we2Bw2CuLgMT4lEiXICIick5CuYf+gLW20RizALgc+DXwP+EtKzLi4uIjXYKIiMg5CSXQ/cHvVwOPWWtfBOLCV5KIiIicrVAC/bAx5pfArcBLxpj4EK+LSgvbc0gMBCJdhoiIyFkJJZhvAV4BFllr64BM+uk8dACPK4F2YyJdhoiIyFkJZZR7C7AXWGSMuRvItda+GvbKIiTWFU+HMbR5WyJdioiISMhCGeX+NeAPQG7w6/fGmL8Pd2FnqCks26cCxLmdgXH1TbW9/toiIiLhEkqX+5eA2dbaB621DwJzgC+Ht6zTC9f2qXAs0Jtb6nr9tUVERMIllEA3HBvpTvDnfnuTOc7tLP/a1NpvV7gVEZF+KJSlX38LvG+MeSb4+Aacuej9UlxMAngV6CIiEl3OGOjW2oeNMSuABcFDt1trPwxrVRHUueNaS1tThCsREREJ3Sm73I0xmZ1fONum/j74VRI81i95Yp313P9l60ORLUREROQsnK6Fvh6wHLtfboPfTfDn4WGsK2KsdT5mTUy/XTtHRET6oVMGurW2+HwW0ldMG30ZVC2LdBkiIiJnRc3Qj5k8ak6kSxARETlrCvRuzPWmMshnz3yiiIhIH6FA70aciaPNpUAXEZHoEco8dIwxbiDv+POttQfDVVSkxZt42vrt0jkiItIfnTHQg+u2fxcoBzr3FbXA5DDWFVFxbg9tLhc+Xzuxsdr6XURE+r5QWuhfA8ZYa6vDXUxf4XEngIXaxkpyMwsiXY6IiMgZhXIP/RAwoNZBjY9JBKCusTLClYiIiIQmlBb6PmCFMeZFwNt50Fr7cNiqijBPbDL4oK5hwHRKiIhIlAsl0A8Gv+KCX/1eYlwKtEBDc1WkSxEREQlJKJuzfA/AGJMcfNzvdy1JjE8FoLlVe6KLiEh0OOM9dGPMRGPMh8A2YJsxZr0xZkL4SzttTdcaYx6rrw/Prf1kTzoAjW0KdBERiQ6hDIp7DPiGtXaYtXYY8E3gV+Et6/Sstc9ba+9KS0sLy+snJTiB3tLeEJbXFxER6W2hBHqStfatzgfW2hVAUtgq6gPSk7MAaPX2+7sLIiLST4Q0yt0Y8wDwu+Djz+KMfO+3UjsD3adAFxGR6BBKC/0OIAf4c/ArJ3is38pIyQWgrnxnhCsREREJTSij3GuBe85DLX1GRko2AJWJFRGuREREJDSnDHRjzCPW2q8bY57HWbv9BNba68JaWQR1rt8eE9BmdCIiEh1O10LvvGf+o/NRSF8zqj2GRpcCXUREosMpA91auz7441Rr7U+Of84Y8zXg7XAWFmkJxNLqaol0GSIiIiEJpQn6hW6OfbGX6+hzPCaONnPSnQYREZE+6XT30JcAnwaKjTHPHfdUClAT7sIiLcF4aDGWNp8fT6w70uWIiIic1unuoa8GjgDZwI+PO94IbA5nUX1BrImn2RjGPfAS+39wbaTLEREROa3T3UMvAUqAueevnL7DTSJtLhcpNEa6FBERkTMKZXOWOcaYtcaYJmNMuzHGb4zp94ucJwYCAIwo+s8IVyIiInJmoQyK+xmwBNgDJAB3Ao+Gs6i+wMY6I9w/SvBHuBIREZEzC2mitbX2I8BtrfVba38LLA5vWZGXkj880iWIiIiELJRAbzHGxAEbjTH/aYy5N8TrotrtCx4CIL0DAgFNXxMRkb4tlGD+HOAG7gaagULgpnAW1RfkJOZwhS8Ng6WxrSPS5YiIiJxWKJuzlAR/bAW+F95y+paUmCSaqKO2xUtaYmykyxERETml0y0ss4VuNmXpZK2dHJaK+pCU2FR8HWVUNDZQlJ0c6XJERERO6XQt9GuC3/8u+L1zs5bPcpqg70/SPOnQBBW1B6F4cKTLEREROaVT3kO31pYEu9uvsNb+o7V2S/DrPuAT56/EkxljrjXGPFZfXx/W90lPdPZFr6k9GNb3ERER6alQBsUZY8z84x7MC/G6sLHWPm+tvSstLS2s75OdMgiATSUfhfV9REREeuqMg+KALwG/McakAQaoBe4Ia1V9RG7GEACO1hyOcCUiIiKnF8oo9/XAlGCgY60Nbz93H5KZVgBAdnJrhCsRERE5vdONcv+stfb3xphvfOw4ANbah8NcW8SlpQ0FYKdbXe4iItK3ne5eeFLwe8opvvq9xCTnHrrpaOdTv1gd4WpERERO7XTbp/4y+H1ALSZzApeL8V6LFx9rD9RGuhoREZFTOl2X+09Pd6G19p7eL6fvSbax1Lu19KuIiPRtpxsUt/68VdGHpboSWBfbAC5vpEsRERE5pdN1uT95Pgvpq4amZBHwNpI0/GEeXzWVOy/UtqoiItL3nHGBGGNMjjHmR8aYl4wxb3Z+nY/i+gJ/bBwArth6/vXFHRGuRkREpHuhrPj2B2AHUIyz29oBYG0Ya+pTCqyJdAkiIiJnFEqgZ1lrfw34rLVvW2vvAC4Nc119xs25sxnrbccTCJAQG9EVb0VERE4plITyBb8fMcZcbYyZBmSGsaY+JXbe17i6qYU2l4tWfwveDn+kSxIRETlJKIH+r8FlX78JfAt4HLg3rFX1JTFx5Iy4HICEIf/LnvKmCBckIiJyslAC/X1rbb21dqu1dqG1doa19rmwV9aH5CQ7e6HHJO3jmv9+J8LViIiInCyUQH/XGPOqMeZLxpiMsFfUB41ILT7hsbU2QpWIiIh074yBbq0dDXwHmACsN8a8YIz5bNgr60Oyhs4jxR/oevzNP26KYDUiIiInC2nYtrX2A2vtN4BZQA0wsBadyRvP7a3BwXDGx583aH90ERHpW0JZWCbVGPMFY8zLwGrgCE6wDyg5+dMBMDENxLo1N11ERPqWUFrom4CpwPettaOttfdZawfcOu+5wYFxiakf4olx6z66iIj0KafbnKXTcKv0IiUY6K7c12msvpzq5nayk+MjXJWIiIgjlEFxAz7MAUblTgYgwTq/skde3x3JckRERE6gtUxD5EnJ59LmFlpNAHfCAX7/3kFa27VqnIiI9A0K9FDljGOjx+liTyz6BQC3PrYmkhWJiIh0CWWU+2hjzBvGmK3Bx5ONMd8Jf2l9jMvFzLhs58cOJ9gPVDVHsiIREZEuobTQfwX8E8FNWqy1m4HbwllUX/VPtQ0AFFpnPffMpLhIliMiItIllEBPtNZ+8LFjHeEopq/LXvAtbq9r4HBMLLddUMDRhjYCAY0ZFBGRyAsl0KuMMSMAC2CMuRlncZmIMcZca4x5rL6+/vy+8ZRPU5g8hA4DQ3LaafMFGP7tl3hs5d7zW4eIiMjHhBLofwf8EhhrjDkMfB34SlirOgNr7fPW2rvS0tLO7xu7XAzLnQjA9pY/dx3+t5d2sq3sPP9xISIicpxQAr3EWns5kAOMtdYusNaWhLmuPmt4zhQAVpW/eMLxGx59NxLliIiIAKEF+n5jzGPAHKApzPX0edn5U8nvcIYQzBzl7Tru81ueWnuQrYfVUhcRkfMvlEAfC7yO0/W+3xjzM2PMgvCW1YdljezaSjUx/zmevGMWeanONLb7/rSFa/77nUhWJyIiA1QoS7+2WGufttZ+EpgGpAJvh72yvio5l0VeJ9Ctbefi0TlcPi7vhFPueGItFY1tkahOREQGqJBWijPGXGyM+TmwHvAAt4S1qj7uzuoK5ra2Ulu5HYB7Lht1wvNv7qxg2QeHIlGaiIgMUKGsFHcAZ2T7KmCStfYWa+2fwl1YX+aa+hkmetspxY8v4CMv1UNaQuwJ5zz82m6m/8trPPL6bjqCXfQiIiLhEkoLfbK19kZr7VJrrdY6BbjqhxQaD34Drx14jYAN8NevX8ifvjqP+68c23VaTXM7j7y+h5+v0Dx1EREJL3Oq3VGNMf9orf1PY8xPu3veWntPWCsLwcyZM+26desi8t4lz/0d19SuBOD2ibfzjRnf6HrOWkvxP710wvlfvrCYG6YVMGHweZ47LyIi/YYxZr21dmZ3z52uhb4j+H39Kb4GtGGDpnb9/Nutvz3hOWMMP7ltKtdOGdx17Fer9nP1T9+h6P4XeWrtQZavL9WysSIi0mtO2ULv9mRjXECytbYhfCWFLpItdA6+zxde/iwbPB4Avjnjm3xx4hdPOq213c+4B/962peaPzKLqYXpXDw6l+lD0+kIWDyx7nBULSIiUex0LfQzBrox5v9wlnr1A2txpq39xFr7w94u9GxFNNCtpez/ZbGosKDr0JYvbOn21LUHavjVyn0crGlh59HGs3qbm6YP4WuXjWJoVmKPyhURkeh3ukCPCeH68dbaBmPMZ4CXgftxutwjHugRZQyDU4ay7HAptxXk4zanvntxQVEmFxRlAlDf4uPdvVVsLq3nF2+febDcnzaU8qcNpdwxv5hvfmI0SfGh/CcTEZGBJpQW+jZgKvB/wM+stW8bYzZZa6ecjwJPJ6ItdICdL8GyJfwkI43H09NYfu1yClMKSYw9u9b0oZoWclLiOVrfxps7K/jhK7to9fm5cFQ2q/ZUnXDuVZMG8cA148lPS+jNTyIiIlGgp13u9wD3AZuAq4GhwO+ttRf2dqFnK+KBDrD2cf6w6rv8IMtpgS8oWMD/XP4/vfoW60tquOl/1pxwbFJBGnOGZ3LR6BxG5iZT3uAlMc7NsKxE4mN0/11EpD/qUaCf4gVjrLUdPa6sh/pEoAO+v9zN9Lpjq+Ge6l56T63cXckXfvsBZ/pPlp0cx6dmFvK5OcMYnK6WvIhIf9HTFvrXgN8CjcDjOOu532+tfbW3Cz1bfSXQWflDJu3/3xMOvXPbO6TFh2fO+e7yRv7rtd28tr2cjjNMfUv1xDCrOIslswq5aHQOse6QVvsVEZE+qKeBvslaO8UYswj4G+AB4HfW2um9X+rZ6TOBXlvCrp/P4JXkRH6V7oT430z+G+6edvd5eXufP0CMy2CMoa6lnec3lfHsxjLWl9SedG5mUhzL7prD6LyU81KbiIj0np4G+mZr7WRjzE+AFdbaZ4wxH1prp4Wj2LPRZwId4CEnyCcVDwXAYNj0+U0YYyJZFTuONLD0g4P875qSE47HuV0smVXIyNxkphZmsLu8kb2VTVw2LpcZwzIjVK2IiJxOTwP9t0ABUAxMAdw4wT6jtws9W30q0D/4Fbz0LXbExfKF/DxaXU7X9jPXPcPIjJERLs7R7O3gnY+q+JvfnXmhv9F5yfzz1eO5aFR2xP8oERERR08D3YUzbW2ftbbOGJMFFFhrN/d+qWenTwV6p4fSqHC7uWzosQVn+lKoA3T4A+w82shLW46csHHMRaNz2FPeyJH6E/dyT4xzs2BkNldOGsTVkwYTF6P78CIikdDTQDfAZ4Dh1trvG2OGAoOstR/0fqlnp08G+pbl8Kcv8fXcbN5Icuajzx40m8cXPR7hwkLX5vOzZl819y3fTEWj96Tn89M8XDM5n5lFmVQ1eYZZcO0AACAASURBVJlVlMmwrKQTgt7b4ae+xcfRhjY6Apac5Hiqm9vJSYnH77ekJsTQ7g+QmRhHjAbqiYiEpKeB/j9AALjUWjvOGJMBvGqtvaD3Sz07fTLQAVY9jPfN7/FOQgJfz8sBYMnYJdx3wX24XdE1R9xaS12Lj1UfVfHmjnK2ljXwUUXTSee5DEwsSOPCUdksX19KecPJfwicSkp8DInxbuYMz2JaYTpzRmQxJi9FXf0iIh/T00DfYK2dfvxAOK0UF4LgILnvZ2Xwx9RjI8rfXfIuqXGpkaqqV1Q0tvH2rkpe3HKEzMQ4mts72FPRxL7K5pPOzU6OoyA9gVi3C7+1DE5LYOXuSi4ek8MLm4+c8j1S4mO4bFwuM4symVqYzoTBqQp4ERnwehro7wPzgLXBYM/BaaFrlPvp7H0LfncDAIuHDOZw7LE12PtDqHfnw4O1rD1Qw4icZKYUppOdHB/ytYGApaalnT3lTazYXcF7e6vZVd5Imy8AQGFmAgtG5rB44iAWjMzG7VK4i8jA09NA/wxwKzAdeBK4GfiOtfaPvV3o2erTgQ5gLXwvHa+BX6andc1RB7hz0p3cPfXuqOuCP59a2jtYtaeKFbsq2F3e1DWvPjs5jpumD2HBqGwmDE4jMykuwpWKiJwf5xzowRHuc4Aa4DLAAG9Ya3eEo9Cz1ecDHaCtHn7gzE2vdxkWDCvsempyzmSeXPwkMS7toBaKykYvq/dW8Yf3D7K+pBZ/cJU8T6yLCYPTmD8ii6sm5zN2UP/r/RARgZ630PvEIjLdiYpAB1j5Q3jzXwGocrtYOHTICU8/e/2zjEgfEYnKolZ1k5c3dlSwrqSGgzUt1Ld2sOtoAwELxdlJzBmeyaDUBKYUpjFneBaeWPWEiEj062mg/whYA/zZnstOLmEUNYEOsH8lLP8SNFcA8ERqCj/Oyuh6en7BfB6c8yCDkwdHqsKoV9HQxrMbD7N8fSn7KptPWOd+dF4yF47K4YapBUws0AA7EYlOPQ30RiAJ6ADacLrdrbU24v2aURXonV78Jqx15qQ/n5TIt3OzT3j6hxf9kMXFiyNRWb8SCFjK6lt5b18Nmw7Vsb6klu1HGgBnHv30YRmMzUvhYE0LuanxzBiWwbwR2WrJi0if1uvbp/YVURno4IyAf+qz0O7M5340PY1fZBwbMDc+azz/d9X/acBcL6toaOO5TWW8ur2cTYfq8HYEiI9x4e1wRtJnJcUxqziT7OR4Lhqdw0Wjs7W3vIj0KT1toXe3q1o9UBLpPdGjNtDBGQH/6ndgzc+6Dv06LYVHMo91w//XJf/F5cMuj0R1/Z61ltLaVrKS4+gIWNYfqOUP75fw+o6KE84bk5fC9GEZLBiZzezhmWQmxuHSlDkRiZCeBvp7OFPWtgQPTQK2AmnAVyO5L3pUB/rx1j4O7/wE6g/y16RE/uG4bvhbRt/CA3MfiGBxA4u1ln1VzRysaWHV7irWldSw80gj7X6nFe92GWYMy+CysbksGJXNuEGpCngROW96Guh/Bh6w1m4LPh4PfB/4R5yBclN7ud6Q9ZtA77T3TfjdjTQbwycL8ikLLkYzKmMUT13zFLGu2AgXODA1ezv4YH8NHx6qY3NpHSXVLeyvclbFi4txMbs4k6GZiYzNT2XcoBTGD04lMU5TEUWk9/U00Ldaayd2d8wYs1GB3suaKuCFe2HnC5TGuLmy0Nm1bXDSYJ65/hkSYxMjXKAA7Kt0FrpZuaeKDSW1HK5rPeH5ETlJTC3MYMawDMblpzA6L4WkeIW8iPRMTwP9KZyFZZYFD90KZAOfA96J5CYt/TLQO7XWwpbl+F76FtOLh3YdfnLxk0zP625Yg0SSt8PPvspmtpTWs/1IA7uONvLhodqupWvdLsOwzEQGpycwsSCNKUPSWDAqmxSPel1EJHQ9DfQE4G+BBcFD7wI/x5nClmitPXnrrfOkXwd6p/Zm7A9H8Y/pHv6anATAA3Me4JYxt0S4MDkTb4efrYcbOFDVzKbSOo7Ut1FW18ru8kZ8fkuMyzA8J4nx+anMHp7F6LwURuYmk5agkBeR7vV42poxJg4YA1hgl7XW17slnpsBEegA/g7s72/gqy07eTcxAYAbRt7AA3MeIM6tdcyjTWu7n3UlNazZW822sgbWHaihud3f9fyo3GSmD82gMDOBSUPSmV2cqfnxIgL0vIV+Cc6mLAdwFpUpBL5grV3Zu2WevQET6ACBAPzhJv5S/j7fycnqOvzt2d/mxpE34onxRLA46YlAwBlZ/1FFE9vL6ll7wFkEp77V+bvZGCjKSmLByGwWjMpm+tAMspI0fU5kIOppoK8HPm2t3RV8PBpYaq2d0euVhsgYcy1w7ciRI7+8Z8+eSJURGaXrqPvNFXyicDCtLtcJTz1y8cNcOuxyLWvaT1Q0tPHe/hq2ldWz+2gj7++voSXYks9P8zBmUApFWUlMHpLG6LwUxgxKIdbtOsOrikg062mgb7bWTj7TsUgYUC3047W3YP/vUxwqfY8783M5EnPi6OnXb36dvKS8CBUn4dLeEWDDwVo2HqpjS2k9Ww7Xc6S+FZ//2P/DOSnxDM9OIiMxjnH5qYzNT2FqYTp5qerBEekPehrovwX8wO+Dhz4DuK21d/RqledgwAZ6J18rrH+Sqlf/iV+np/L7tGPL69+WN5d/uPQR4uI0za0/a+8IsL+qmZ1HG9h4qI6qpnaO1LVS3dzOgepmOv/3HpGTxMSCNBaOyWVkbjJF2UkkaxqdSNTpaaDHA3/HsVHuq4CfW2u9vVrlORjwgX68QAC7/S88tOIb/DklGYAYa3ly3r8zefS1ES5OIqGxzdfVkv9gfw0bD9VR3dwOONPoBqV6GDsohbkjspgzPIuhWYmkahqdSJ92zoFujHED26y1Y8NVXE8o0Lvh76Bu61P82zsP8nKy0zof7OvgN5f+jILhl0W4OImkDn+ADw/Vsbu8kZLqFg7VtLDraCP7gqveGQOTh6QzY2gGY/NTmDwkjTF5KRqTIdKH9LSF/hfg7621B8NRXE8o0E9v7/Y/seSD79Ia/Af5cX8Ws5c8AwkZZ7hSBpIj9a28v6+GnUcbeX9/NdsON3StXZ/iiWF0nnMfflJBGiNykhmRm6SlbUUipKeBvhKYBnwANHcet9Ze15tFngsF+pnVNh7lb178NDu8lQAMb/fxaNY8hnzyNxGuTPoqf8Cyv6qZ9SU1bD3cwObD9ew80tC1zSzA4DQPEwvSmFiQxtTCdCYWpJGRGKvWvEiY9TTQL+7uuLX27V6orUcU6KHbV7Wdr79yJ/s7GgF4mgLGXf5vUKBlZOXMvB1+9pQ3sbeyidLaVnYdbWTr4fqu7nqApDg3+ekJjMlzNqgZnO6hODuZYZmJpCXEYnHu3YvIuTunQDfGeICvACNxtk79daT3P/84BfrZsdbyXskb3PX2vQAsaGnl9voGZo79FK7LHoCUQRGuUKJNQ5uP9SW1fFTeRGltCwdrWthyuIGqpu7HzKYlxJKVHEdRVhLZyXFkJceTkxxPXqqHQWnxjMpL0cA8kdM410B/CvDhjGq/Eiix1n4tbFWeAwX6uSlpKOGfX/97NjXu7zr26fpGrmtqYkLeDLj1d5CcG8EKJdo1tvn4qKKJA9XNHK33UtPsxeUyNLT6OFrfxtEGL2V1rV2r4R0vLzWezKR4clPiyU/zMDg9gaGZiYzMTaY4O0m71smAdq6BvsVaOyn4cwzwgbW2T/XPKtDPnbWWdUfXcserXzrh+PB2H4+WVzCkcD4seQo0j13CxFqLtyNAdXM7tc3tHKxpYV9lE3srm6lraXfm1Ne3Ud3s5fh/ptISYklPjCUjMY7h2UkUZCRQkJ5ATko82cnxpCbEMjQzUd370i+da6BvOD7AP/64L1Cg946ShhJWlq7klf1/ZVPVZgBur2vga7V1uIfMgmt/AnnjI1ylDFRN3g4O1bRQUt3M3spmjtS3Utvso7LRS2ltC0cb2gh87J+xhFg3Y/NTGJ+fyvjBqYwdlEJhRiI5KfEauCdR7VwD3c+xUe0GSABagj9ba21qtxeeRwr03rerZhffevtbHGg4gMvCP1fXcEtjE8SlwLWPwKSbI12iyAl8/gBH69vYV+W07L2+ADuONrC9rIHtRxpobDs29CfFE8P4/FSyU5x798XZSRRnJzEqL5lBqR6FvfR5Pd4+ta9SoIeHP+Dn0Y2P8qstvwLAWMv/q6zmmuYWTM5YuOJfYORl4NKWntK3WWsprW1lT0Ujh2qcvei3lTVQ0dBGQ1sHTd5jYZ+RGMvI3GQK0hMYnJ7A6LwU0hJiKcpOYkhGgja+kT5BgS7npKq1irvfuJtt1dsAcGP4YWU1VzQ1OSfMvRuGzoVRV0BMfAQrFTl71loqm7zsq2xmd3kj28sa2FvZRFldG+UNbXQc14/vdhkmDE5lYoGzet6ovGRG5iSTnRyvbWzlvFKgS480tDfwX+v/i+W7lwNQ6ErgW0cOcmlLq3OCJw0uuBPmfx08Eb8TI9JjPn+AA1XNVDZ5Ka1tZX9VMxtKatl5tPGEkfkuAymeWDKT4nC7DFlJcWQkxpGRFEtCbAxJ8W5i3S7iYpzWfV2Ljw5/AG9HAG+HnzZfgPTEWBLjYqhobKPZ20FCrJu4GBcZiXHkpDhT+nKD9/4T49zEB18rIymOzMQ4jEG3CgYQBbr0iraONn615Vf8dutv8QWcf9SuTBzG3x4poagmuDLwgnth0CRoa4DETCiYCUk5EBMXwcpFeoe1lspGL7vKGzlQ1eyMwm9qp7LJiwEa2zqoaWmnstFLa7u/awnd47kMxLpdeGLddP7729DmBHmyJ4b4GBcNrT7afIFur/+4GJdhUJqHrGRnql9yfAwZiXFdr1WYmUiqJ4aRucnEx7jJSopTr0IUU6BLrzrSdIQfrvshr5W81nVscvJQ/mXnewz3nWLtofwpkD4Uhl8Ck2+D+OTzUqtIJHX4A3QELIHgv7PxMW5c3bSoO/8dPv64tZaG1g6ONrRxtKGNFm8HLe1+fP4AvoDF7w9wpL4Nn99ysKaF+lZnml+bz0+z10+rz99tTTEuQ6zbRW5qPCmeGAalJpCf5mFQmofi7CTaOwJkJccxOi+lq2dA+g4FuoTF0eajBGyA32z9DU/vehqLZZQnh8+7cxnV4WeYdZPc4YW9b558ccEMp+U+/QswejG4NOBIpLd0/jHgdht2HW2ksc3HgapmfH5LbUs7h+taaWn3U9HQRl2rj5Lqlm5fJynOTUKcm/y0BLKT48hOjmd4TjKD0z1kJcWTmeTcFkhNiKHDb0mMc+sPgDBToEvYlTSU8M0V32RX7a4TjucmOivOfXvmPzI/NhNPyRpY/wQ0V4G33jkpMQvSh8GCr8OoRRDrOc/ViwxsgYCl1ednf1Uz9a0+rIV9VU3sOtrIwZoWjDHUNrdT3tBGRWP3y/oCpCfGUpydxLj8VDISY8lMiicvNZ6hmYkMzUwkPVG33npKgS7nzebKzTy39zlWlq7kSPORk56fmTeT0Rmj+fy4z1Lg98PuV2Hrcjj0vnOCccG0z8HUz0DhLGeTbhHpM+pbfVQEg72y0Utjm4/6Vh/+ABxtaGNfZRM7jjTQ6O3g4/GSnRzP6DxndsDg4Op+mUnOyn4F6YnkpmjWwJko0CWidtXs4oltT7CzZieHmw7T2uGMjk+MSeT6kddz56Q7yY1Jhp0vwtY/wZ5XwfqhcA5MuQ2mLFGrXSTKBAKW+lYfRxvaOBhc6W9PeRN7KpqobvZyNHj//3hxbheD0z0MyUgkxRMT7O73UJiRyKA0D+mJcXT4AxgD6YlxZCXFkZYwsLbtVaBLn9Hub+el/S+xs2Ynz+99nob2BgAKUwq5dcytXD/ietLbGuHD3ztfDaXOhTO+CLO/CrljI1e8iPSaQMBZB6C+1cfhulZKa1sprW3hcK3zc2Wjl4C1VDR68X98bd/jJMS6KchIYEhGgtPyT/OQn55AqieW3FSnJyArKQ5PbP9YCEuBLn3W9urtPL3raZ7b+9yxqXDFV/L58Z9nQvpozJ5XYeMf4KPXwd/u3GO/6j8hoyiyhYvIedHhDzgj/evbqGtx/o2IcRvqWnxUNTkt/UO1LRyua6W6ybnP313+ZybFkZ0cx+D0BAaleijMTGRYViK5KR4yEmNJT4zrWk+gL1OgS5/nD/hZW76WxzY/xpbKLbT525iSM4XPjPsMlw29jLi2Bnjnv+CDxyDQAeNvgPn3wOBpkS5dRPqQ9o4AVU1eGtp8lAe36a1u8lJW30ZVo5fDda2UN7RR1dR+0rVulyEn2RnIl5fqCX4d/7OHQakeUhNiItbNr0CXqFLRUsEL+17g6V1Pc7jpMAD3zriX28bcRmJLDaz4AWxa6gR75ggYexVMvEnhLiIh69zFr6rJS22Lj9pmZ0Gg8uC8/4oGL+WNx3oFjueJdTkBn+IhNzWeQcGwz0yKIz0xlpyUeHJTPGQlx/X6HgAKdIlKHYEOVpet5h/e/gdaOlrI9GRy3Yjr+NToTzHU5YENT8K2v0D5FueCghlw/c91n11Eek2bz98V7kfrnXX+nS9v189HG9po83W/qt/Vk/N59NO9t/O4Al2imrWWdeXr+N3237GydCUBG2B2/mw+N/5zXFhwIaahDLY/C6t+DO3NzmI1Ez8JQy7QjnAiEnbWWhraOqhtbqeu1UdlcEpfRWMbhRmJ3DRjSK+9lwJd+o2jzUf5w44/8Nze56hpq2FoylBuGXMLlxZeSqErHl75Z9j2Z6c7PiUfpn/emdeeXhjp0kVEekyBLv1Oa0crL+9/mT/t/hObqzYDMDt/NteNuI6rcy/AvW8lfPBLOLzeWaymaIEzn33CJzWnXUSilgJd+rVt1dt4o+QNlu5cSpOviYLkAhYWLmRR0SKmEu/MZ9+0DFqqIDkPZt3lDKLLLI506SIiZ0WBLgOCtZY3Dr7BU7ueYn35enwBHyPTR3LbmNu4qvhKUkreg3cehoNrwLhh6Bxwx0H+ZBh5ORRdqKVmRaRPU6DLgFPvref5vc/z9O6n2V+/n+TYZBYXL+az4z7LCK/XWWJ22zNOgFd/5Fw0aDLMvB3GXQdJ2ZH9ACIi3VCgy4BlrWVDxQae2vkUrx18DX/AzyWFl3Db2NuYmz/XWRzC2whb/wxrHoWq4G5xQ+c6y82OuQo8qRH9DCIinRToIkBVaxVPbnuSp3Y9RWtHK0WpRczIm8GUnClcNOQisuLTYd9b8NEbzkYxdSXOhUPnwbTPOAPq4hIj+yFEZEBToIscp7Wjlb/u/yvLdi1jb91evH4vca44rh95PV+e9GXyk/PB3+Hca9/1Eux4AeoPOhcXXQiTb4UJN0J8cmQ/iIgMOAp0kVMI2ACbKjfx3N7neO6j52gPtDMlZwrXDr+W60ZeR0JMAgT8cGAVbFkO+9+GuoPgioGRVzjbu45erKlwInJeKNBFQnCk6Qi/3fZb3jz4JuUt5STEJHBJ4SUsKlrExUMuJsYVA9bCwfdg5wtOwDcdBU+aM0p+2DwYew2kDIr0RxGRfkqBLnIWOgfSPb/3ed44+AZ13jryEvO4bOhlfHLUJxmTOcY5MeCHfStg81Ow720n3DFQMN0ZTDfmKsgbH8mPIiL9jAJd5Bz5/D5Wlq7kj7v/yLtl7wIwLnMcN4++mcXFi0mNC46ADwTg6CbY+ZLTeq/Y7hzPHgNjr3a65YfOjtCnEJH+QoEu0gvqvfU8s+cZnvnoGfbV78Pj9nDFsCu4ZsQ1zB40G/fxG8E0lDld8tufhbIPwQYgPg2KL4Tii2DSpyAxM3IfRkSikgJdpBdZa9lWvY3lu5fz6oFXafQ1khCTwFXFV3Hp0EuZN3iec7+9U0uNs4jN4Q3OoLr6Q85KdcUXwvjrYfwNCncRCYkCXSRMvH4vbx96mzcPvckbJW/Q5m8j05PJRUMu4saRNzItd5qzeM3xjm5xFrLZ/heo2QsYKJwFwxfCiEude/Du2Ih8HhHp2xToIudBU3sT75a9yysHXuHtQ2/THmgnPymf2fmzWVS0iLn5c0/slgenO373q8589yObAOt0zY9Y6IycH70IknMj8nlEpO9RoIucZ03tTbxW8horDq1g1eFV+AI+XMbFjLwZXDb0Mq4qvooMT8aJFzVXw/4VsOd12PMKtFSDKxYKZ8OIS5xR87njtYGMyACmQBeJoGZfM6tKV7GufB0rS1dypPkIMa4YFhQs4Jrh13BJ4SXEu+NPvMhap8W+dbmzUl3tfud4RrGzr/vYqyFvgsJdZIBRoIv0ITuqd/Dc3ud4cd+L1HprSYlN4dKhl7JgyALm5s8lLT7t5IvqD8Puvzqj5vevdI6lD3O2gB1xKYz6hAbWiQwACnSRPsgX8LGmbA0v73+Zt0vfprG9EbdxMy13GgsLF7KoaBF5SXknX9h41Nk8Zt9bcOBdaK1xRs3nT3FWqyu+CPKnQko314pIVFOgi/RxHYEOtlZtZWXpSt44+Ab76vcBcMGgC7hg0AUsGLyA8VnjTx5U5/dB6TrnnvveN+HIZiD4/3TmCGdBm3HXQOEccLnO74cSkV6nQBeJMvvq9/HKgVf46/6/doV7XmIeVxZfyTXDr2F0xuiTp8MB+Frh0AfO6Pn9K51NZfztkFoA466DyZ+CwdN1710kSinQRaLYoYZDrDmyhjcPvsl7R97Db/0UphSysHAhc/LnMCt/1smD6jp5m2DXy7D1T04L3u+F2CQYNtdpvRdfBNmjFfAiUUKBLtJP1LbV8lrJa7x56E0+OPJB13S46bnTGZs5lqm5U5mRN4PshOyTL25rgG1/dlrv+1ZA7QHneEYxTLjBCfghs9Q1L9KHKdBF+qGm9iY2VGzg3cPv8v6R99lbv7frueFpw5meN525+XOZXzCfpNikEy+21gn0vW84G8rsWwHWD6lDnHvuwy9xWu9xH7tORCJKgS4yAHj9XjZWbOTDig/ZWLmRjRUbafY1E+OKYWHhQj4x7BMsHLqw++75lhr46A3Y8kdnvfmONnDHOaPmR14OxRfDoEnqmheJMAW6yADk8/vYVLmJ1w++zkv7Xuqa8z45ZzIzB81kZt5MJmRPINb1sXXjO7xwcA3seQ32vApVu53jaYXO1Lihc50WvBa2ETnvFOgiA1zABvjg6Af8ec+f2Vy5mcNNhwFIjk1mfsF8Li28lIsLLz65ax6g4YgT7HtedUbQN1c4x1MGw5jFzqI2RRdCfPJ5/EQiA5MCXUROUNtWy9qja3nn8DusOryKqtYq4t3xXFJ4CZcPvZz5BfNJiUvp/uK6g7D3LfjoNfjoTfA1AwayRzmD6govcOa9Z4/WADuRXqZAF5FTCtgAH1Z8yMv7X+aVA69Q563DZVxMzp7MgoIFXFx4MaMzRuMy3YRzhxdK3oWD78PB1c7WsK21znMJGU73/LD5kDMGUgY5IR9ziil2InJGCnQRCYk/4GdT5SZWl61mZelKdtbsxGK79nifmDWRaXnTGJE24uRV68AZPV+9Fw6959yH378K6kqOOyHYks+bCPmTne85Y5z787ofL3JGCnQROSeVLZWsObKGVaWreLv0bVo7WgHI9GRywaALuHjIxVw05KLuN5TpVF/qtNybyqFmH1R95DyuP3jsnIQMp/U+bL6zF3zeRG02I9INBbqI9Ji1lj11e9hYsZENFRt4/8j7VLVW4TZuJmRPYErOFCZkTaAotYihqUNPfQ++U2O5s8hN7QGo3AlHN0PZRmc+PEDaUMgbDzljnV3l8iZC2hC15GVAU6CLSK8L2ADbqrbx1qG3+ODoB2yt2oo/GMYGw8iMkUzLmcbMQTMpTismIz6D3MTc7teg7+RthJI1TsAfXu+05Gv3gw04z6cWODvJ5U859pUySCEvA4YCXUTCzh/ws6NmB+vL19Pka2Jz5WY2V26mydfUdU6mJ5Oi1CLGZ41nQvYERqWPYkT6CGJcMad+YV+r05I/utW5L1++Far20LWrXFKuE+w5YyA5F3InwJCZkJAe3g8sEgEKdBGJCH/Az7bqbeyp3UNDewP76vdxoP4AO2t20uZvA5zWfH5SPmMyxzAjbwbzBs9jWOow4txxp35hb5MT7Ec2Hfuq2HGsux4go8jpph88DQqmO98TMsL7gUXCLCoC3RgzHPhnIM1ae3Mo1yjQRaKTL+Bjd81udtbs5EjzEQ40HGBjxUbKW8oBcBs3I9JHMDZzLEWpRYzJHMPk7Mmke07T6rYW2uqcbvpD7zv34yt2QM2xNe5JGey05uOSnMDPnwzJg5xV77QwjkSBiAW6MeY3wDVAhbV24nHHFwM/AdzA49baHxz33HIFusjAY63laPNR3i17l7KmMrZVb2NnzU5q2mq6zhmeNpxpudMYlzmOEekjKEorIiM+o/spdJ1aa51wL9vgdNuXrnVG3hvXsRa9cUHOuGArPh1yxzmD8eKSIXM4xJymt0DkPIpkoF8ENAH/2xnoxhg3sBu4AigF1gJLrLXbg88r0EWkS21bLbtqd7G1aisbyjewsXIjje2NXc8nxyYzPms8U3OnUpRaxPC04YzKGHXqLvvOf/N8rU63fW0JVO+B0nXOQLz2Zgj4jp3vioXMYiiY6ayCN3i6s1HN6f6IEAmTiHa5G2OKgBeOC/S5wEPW2kXBx/8EYK399+Dj0wa6MeYu4C6AoUOHzigpKTnVqSLSD1lrKW8p56O6j9hXt4+DjQf5sOJDdtfu7jon1hXL2MyxTMmZwsxBM5mQNYG8xLzTj7DvFAg48+UrtjnhXrnL2aDm0AfQUuWcE5cCOaOde/KBDojxOMezRjpd+f52p9XvSccZvGcgJc953f/f3r0Hx3mVdxz/PpJWV+u+si35KlvyRZKN7VxIQktdYEgIpKGdlKSlEKCUoZ3p0LRMB4ZOGdphmNIO5dZCaaDhUgI0UApkOpBCmECA3IhjyZItyZJtRb5pJXklW9bF1ukfhfsxnQAAEuBJREFU5+xq7diKZcva1er3mXln3/e877579vi1nj3nPe85Y0NwJuZbCUZe9OewHJ82ecbPdDdxGgrLfEtBSQ1EiiEnz/fmH4/7HyPlq6Gg1M+KV1jmz1NY7jsJVjdAbgSmz6t1IctkWkC/B7jDOffusP024JXAh4GP4mvuDyYC/GxUQxeRhPhEnBNjJ+iN97Ivto/WWCutsVYmzk8AUJpfSmNFI42VjWwo30BDRQNbq7e+/PPyCYk55F981o+EF+uCkX7/SN30eR/UTx32AflK5ESgJArLVvgAPdgNa272zfyW4wPyUK9/hC+1sx/4Y85P+uWSjOQPCRyU1vr8RxuheiOMj8DyJn87om4n5OZBZb1/UiAn4lsf9ChgRpotoM/yrMjCcs4NAu9Ndz5EZHEqLyinvKCcTZWbuH397QBMnp9k3+A+DgwdoGu4i65TXTza8+gFj9IV5RWxLbqNLVVb2LF8B+vK1lFfVk8k96JpZc1803tVPWz//Utn4vw5GD3mg3xuxNemnfPj148eg3g/rLvVB9hEjfvlnAtBOycXMP8eM/8jYvSY/yFxsh3OT0GkCEaO+h8HA/thcsy/b/SYD+JDvX4An/E47PvO5T+zuNoH/Iq1vkVh+dYw2U6uHwegukET72SgjGtynwvV0EVkrpxznBw7SfepbtoH2+k/3U/7YDs98Z5kbT6SE6GxspHKwko2VWyiKdrEjpodrCxZmebczwPnYGIELNe3OLjz0P9r33wf64LRozA17vsVxLohpb/CDPOtCwVlULHG1+qLKn0HwuoG/wOgsBzK6tTXYJ5lWg39GaDRzOqBfuA+4A/TkA8RWYLMjBUlK1hRsoJXrXpVMn3q/BStsVb6T/fTOdxJ53AnsbEYTx97mqnQSa6upI5dK3Zx44obWV++PjkC3hXdm88UZj7YAqwMDx/VvuLSxzrnZ9Qz8zX+WJffHurxI/idPunv9w92+0cGk039QW6+H91v2XJYud037yf6BWgY33l3vXu5PwzsBqLACeDDzrkvmtmdwCfxj619yTn30as5v2roInK9TZ2fonO4kz0De3juxHM8d+K5Cx6ly8/JpyRSAkBDZQMNFQ1UF1ZTlFfEsvxlFOQWUFFQwdqytURyIsQn4uRaLjk5OZybPkdtSe2V38fPVIk4cn7SB/djL/h+AcOH/I+A0yd8x8LJmVsdFJT5VoFIsa/VV2/0HQor1vl7/fklvlOhOvVdYFEMLHM1FNBFZKE55zg8cpgjo0foG+3jxJkTxM763u+HRw7TE++54B79lSjMLUy2ApRESqgtqWV16Wrqy+vZXLWZsvwyKgsqqSqsojS/lEhO5KX3+DPd9Hkf7E+0+c54J/f7bTfte/gPHXxph0LL8TX5qo2+Ob9y3Uy/hNJa/wMg2ujv9S+Rpn0FdBGRBTQ1PcXY1BjxiThnz51ldHKUvtE+4hNxygrKKMgtYGh8iEhOhNHJUY6eOcqyyDJOTZxi2k0Tn4jTN9rHkZEjnHPnLvkZq5atYkP5BpYXL6e+vJ66ZXVsqtxESaSE6sLqxXUbAPzjgqeP+1r9UI/v0DcW8+uDB33AH4/7Yy33wp7/eYW+lh9thOjm8LrJp+UXp+XrXC8K6CIii9D4uXE6hzsZGBvgzLkzDJ4dZOzcGJ1DnRTkFtA+1M7JsZPJeeoTKgoqks/iF+YVsqNmBzesuIGygjJWFK+YfZz8TDY+4p8eyCv09++HenxTfqzTN+3HOv2jg4nZ+cBPw1uzyQf4RKCPbvL38Rfbjx4U0EVEspZzjqNnjnLizAl64j3EJ+IcHjnMnoE9HD19NNlzP8EwaopqWFmyktWlq2moaGBN2ZrkKHuLNtgnTI372nxqkE+sT43NHFdYMRPcE4G+ZrO/h5+bMU90v4QCuojIEjU1PcXQ2SH2DOxhbGqM42eO03+6nyOjR9gX28fk9MzgNHmWR3lBOatLV9MSbaG5upnivGLWlq2lvrx+9mluM930tB8I6CWBvtN32kvIifgOetFNfjz/5VtmBt3JL0lf/oOsC+hmdhdwV0NDw590dXWlOzsiIotWfCKebNY/MHwgWcvfP7T/gqb8wtxCokVRGiobaKpqYkPFBmqKalhbtnZx3rNPdfaU76AX6wxD/Xb5x/SGey9svi9bNfOsfc1mH/CjjX4WvwUaaCfrAnqCaugiItfHuelz9MR76BvtY2xqjI6hDvpG+jg0cojDI4dxKc+bF+UVUVlQyabKTbREW2iJtrC+fD11JXWLO9BPjftAP9TjA3yig17swEwHPbiwU17NFr9esdZvF1XOa5YU0EVEZN6MTY1xZPQIg2cHOTRyiBdHXyR2NkbncCc98Z7kcZUFlTRFm2ip9s33LdEWaopr0pjzeeKc75Q3sN/frx88ONN8P3yYCwbX2XoX3Pu1eftoBXQREVkQo5OjtA+20xvvpX2wnbbBNg6eOsh0aLpeXryclmpfi2+qbmJt2VpWL1u9uGvyqRID6gwf9gG+tPbyY/9fBQV0ERFJm7GpseSc9m2xNvYN7uPwyMzU15UFlbREW9gW3ZZ8rSisSGOOM1emjeUuIiJLSHGkmJ3Ld7Jz+c5kWnwiTvtgO32jfbTF2miNtfLz/p8n782vKV1DS7SF7dHttERbks/Uy+Wphi4iIhnhzNQZ2gfbaY21JoP88TPHAf9IXWNlIxsrNvphccvqaY42s7Z0bfY0118BNbmLiMiiNDA2kAzwe2N76Y33EjsbS96TL8svSzbVb6/ZTnN1M9VF1WnO9fWjgC4iIllj8vwkvfHeZC1+b2zvJTveba/ZngzyxZHsGNNdAV1ERLJa4ln5tlib710fa+PI6BEAciyHjRUbaaluYXPVZpqrmxftPXkFdBERWXKGx4dpjbX6ZaCVjqGO5Fz2eZbH5qrNbK/ZzrboNpqqm1hXti7jh7fNuoCuoV9FRGSunHMMnB3w9+MH9rI3tpe2WFtyiNvC3EK2Vm9lW3Qb22q2sT26ndqS2ozqdJd1AT1BNXQREbkWiSFuDwwdSDbVdwx1JGepK42UsrlqM03VTTRXN7Mtuo3VpekbCEcBXURE5ApNTU/ROdxJ60Ar3ae66Rjs4MDwgWSQryiouGAgnIXsWa+BZURERK5QJCdCc3UzzdXNybSp6SkOnjqYvB/fGmvlyf4nkwPh1JXU0RxtTg5pu7VqK+UF5Quab9XQRURErkJiIJx9sX20DfphbftP9yf315XUcUf9HTxwwwPz9pmqoYuIiMyzkkgJN628iZtW3pRMGx4fpmOog/1D++kY7KAwd+EejVNAFxERmSeVhZXcVncbt9XdtuCfnbPgnygiIiLzTgFdREQkCyigi4iIZAEFdBERkSyggC4iIpIFFNBFRESywKIM6GZ2l5l9IR6PpzsrIiIiGWFRBnTn3Pedc+8pL1/YYfVEREQy1aIM6CIiInIhBXQREZEsoIAuIiKSBRTQRUREsoACuoiISBZY1POhm9kAcHgeTxkFYvN4vqVIZXjtVIbXTmU4P1SO126+y3Cdc67mUjsWdUCfb2b27OUmjpcrozK8dirDa6cynB8qx2u3kGWoJncREZEsoIAuIiKSBRTQL/SFdGcgC6gMr53K8NqpDOeHyvHaLVgZ6h66iIhIFlANXUREJAsooANmdoeZHTCzbjP7QLrzk0nMbI2ZPW5m7Wa2z8zeF9KrzOwxM+sKr5Uh3czs06Es95rZrpRz3R+O7zKz+9P1ndLFzHLN7Hkz+0HYrjezp0JZfdPM8kN6QdjuDvvXp5zjgyH9gJndnp5vkj5mVmFmj5jZfjPrMLNbdS3OjZk9EP4vt5nZw2ZWqGtxdmb2JTM7aWZtKWnzdt2Z2Q1m1hre82kzs6vKqHNuSS9ALnAQ2ADkAy8ATenOV6YsQC2wK6yXAp1AE/Bx4AMh/QPAP4T1O4H/BQy4BXgqpFcBPeG1MqxXpvv7LXBZ/iXwdeAHYftbwH1h/fPAn4b1PwM+H9bvA74Z1pvC9VkA1IfrNjfd32uBy/DLwLvDej5QoWtxTuW3CugFilKuwXfoWnzZcns1sAtoS0mbt+sOeDoca+G9b7iafKqGDjcD3c65HufcJPAN4O405yljOOeOOed+HdZHgQ78H4W78X9cCa9vDut3A19x3q+ACjOrBW4HHnPODTnnhoHHgDsW8KuklZmtBt4IPBi2DXgN8Eg45OIyTJTtI8Brw/F3A99wzk0453qBbvz1uySYWTn+D+sXAZxzk865U+hanKs8oMjM8oBi4Bi6FmflnHsCGLooeV6uu7CvzDn3K+ej+1dSzjUnCug+OPWlbL8Y0uQiobltJ/AUsMI5dyzsOg6sCOuXK8+lXs6fBP4amA7b1cAp59y5sJ1aHsmyCvvj4filXob1wADwH+HWxYNmVoKuxSvmnOsH/gk4gg/kceA5dC1ejfm67laF9YvT50wBXa6ImS0Dvg38hXNuJHVf+FWpxyUuw8zeBJx0zj2X7rwscnn4Zs/POed2AmfwTZ1JuhZnF+7z3o3/cVQHlLC0Wieui0y57hTQoR9Yk7K9OqRJYGYRfDD/T+fcd0LyidBURHg9GdIvV55LuZxfBfyOmR3C39J5DfApfFNcXjgmtTySZRX2lwODLO0yBF9zedE591TYfgQf4HUtXrnXAb3OuQHn3BTwHfz1qWtx7ubruusP6xenz5kCOjwDNIZenvn4jh/fS3OeMka4X/ZFoMM594mUXd8DEr007wf+JyX97aGn5y1APDRL/RB4vZlVhlrC60Na1nPOfdA5t9o5tx5/ff3EOfdW4HHgnnDYxWWYKNt7wvEupN8Xeh7XA434zjRLgnPuONBnZptD0muBdnQtzsUR4BYzKw7/txNlqGtx7ublugv7RszslvBv8vaUc81NunsPZsKC75XYie+p+aF05yeTFuA38E1Je4E9YbkTfx/tx0AX8H9AVTjegH8JZdkK3JhyrnfhO890A+9M93dLU3nuZqaX+wb8H8Fu4L+AgpBeGLa7w/4NKe//UCjbA1xlT9jFvAA7gGfD9fhdfG9hXYtzK8OPAPuBNuCr+J7quhZnL7OH8X0OpvAtRX88n9cdcGP49zgIfJYw6NtcF40UJyIikgXU5C4iIpIFFNBFRESygAK6iIhIFlBAFxERyQIK6CIiIllAAV0kQ5jZx8zst83szWb2wTm+tybMhvW8mf3mLMfttjDb2yzH7DCzO+fy+QvNzA6ZWTTd+RDJJAroIpnjlcCvgN8Cnpjje18LtDrndjrnfnaN+diBH2tARBYRBXSRNDOzfzSzvcBNwC+BdwOfM7O/vcSx683sJ2Ge5R+b2Voz24GfyvFuM9tjZkUXvecO8/OH/xr4vZT0m83sl6FW/wsz2xxGS/w74N5wrnsvddwl8lVrZk+E97QlWgnM7HNm9qz5+bc/knL8odAisSfs32VmPzSzg2b23nDM7nDOR83Puf15M3vJ3ywz+yMzezqc69/Mzzufa2YPhby0mtkDV/WPI7KYpHsEHi1atDjwwfwzQAR4cpbjvg/cH9bfBXw3rL8D+Owlji/Ez/DUiB/B6lvMjFRXBuSF9dcB377UuS533EWf81eEURaBXKA0rFelpP0U2B62DzEz5/Y/40d+KwVqgBMhfTcwjh/FLBc/3eQ9Ke+PAltDmURC+r/ih868AT9VZSJ/Fen+N9ai5XovicH4RSS9dgEvAFvwc85fzq3M1LK/iq+Zz2YLfjKOLgAz+xrwnrCvHPiymTXih/eNXOYcV3LcM8CXzE/k813n3J6Q/hYzew9+prRaoAkfvGFmzoRWYJlzbhQYNbMJM6sI+552zvWEvD+MH4o4MW83+FsNNwDP+GGwKcJPkvF9YIOZfQZ4FPjRLGUkkhUU0EXSKDSXP4SfYSkGFPtk2wPc6pw7ex0//u+Bx51zv2t+rvufXu1xzrknzOzVwBuBh8zsE8DPgPcDNznnhs3sIXyLQcJEeJ1OWU9sJ/42XTw29cXbBnzZOfeSToRm9grgduC9wFvwLRoiWUv30EXSyDm3xzm3Az85UBPwE+B259yOywTzX+BnbAN4Kz5ozmY/sN7MNobtP0jZV87MNI3vSEkfxTd/v9xxSWa2Dt9U/u/Ag/gWhzL8nOVxM1sBvOFl8nopN5ufCTEHuBf4+UX7fwzcY2bLQz6qzGxd6AGf45z7NvA3IT8iWU0BXSTNzKwGGHbOTQNbnHPtsxz+58A7Qye6twHvm+3czrlxfBP7o6FT3MmU3R8HPmZmz3Nha93jQFOiU9wsx6XaDbwQjrkX+JRz7gXgefyPiq8DT86W18t4Bj/7VAfQC/z3Rd+vHR+wfxTK5DF80/4q4KehpeNrwJweAxRZjDTbmohkJDPbDbzfOfemdOdFZDFQDV1ERCQLqIYuIiKSBVRDFxERyQIK6CIiIllAAV1ERCQLKKCLiIhkAQV0ERGRLKCALiIikgX+H/lJboJZbH5nAAAAAElFTkSuQmCC\n" }, "metadata": { "needs_background": "light" @@ -356,7 +383,7 @@ }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 11, "metadata": { "tags": [] }, @@ -364,15 +391,24 @@ { "output_type": "stream", "name": "stderr", - "text": "Seed namespaces (singletons and interactions): ['e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nCreated challengers from champion |supervised||classic\nNew challenger size 37, ['|supervised|gi|classic', '|supervised|eh|classic', '|supervised|ad|classic', '|supervised|gh|classic', '|supervised|bc|classic', '|supervised|bd|classic', '|supervised|ae|classic', '|supervised|dg|classic', '|supervised|ei|classic', '|supervised|df|classic', '|supervised|fh|classic', '|supervised|ac|classic', '|supervised|ab|classic', '|supervised|cg|classic', '|supervised|hi|classic', '|supervised|fg|classic', '|supervised|bi|classic', '|supervised|be|classic', '|supervised|de|classic', '|supervised|ci|classic', '|supervised|fi|classic', '|supervised|cd|classic', '|supervised|af|classic', '|supervised|ce|classic', '|supervised|di|classic', '|supervised|bf|classic', '|supervised|ai|classic', '|supervised|bh|classic', '|supervised|ag|classic', '|supervised|bg|classic', '|supervised|eg|classic', '|supervised|ah|classic', '|supervised|cf|classic', '|supervised|dh|classic', '|supervised|ef|classic', '|supervised|ch|classic', '|supervised||classic']\nOnline learning for 10000 steps...\nSeed namespaces (singletons and interactions): ['cf', 'e', 'g', 'b', 'd', 'i', 'h', 'a', 'f', 'c']\nCreated challengers from champion |supervised|cf|classic\nNew challenger size 43, ['|supervised|bg_cf|classic', '|supervised|cf_dg|classic', '|supervised|ab_cf|classic', '|supervised|bh_cf|classic', '|supervised|cf_eg|classic', '|supervised|cf_ef|classic', '|supervised|be_cf|classic', '|supervised|cf_di|classic', '|supervised|cf_ci|classic', '|supervised|bd_cf|classic', '|supervised|cf_fi|classic', '|supervised|bf_cf|classic', '|supervised|ah_cf|classic', '|supervised|ac_cf|classic', '|supervised|ce_cf|classic', '|supervised|cf|classic', '|supervised|cf_cfg|classic', '|supervised|cf_gi|classic', '|supervised|ag_cf|classic', '|supervised|ae_cf|classic', '|supervised|cf_fg|classic', '|supervised|cf_hi|classic', '|supervised|cf_df|classic', '|supervised|cef_cf|classic', '|supervised|cdf_cf|classic', '|supervised|cd_cf|classic', '|supervised|bc_cf|classic', '|supervised|cf_gh|classic', '|supervised|cf_cg|classic', '|supervised|cf_ch|classic', '|supervised|bcf_cf|classic', '|supervised|af_cf|classic', '|supervised|cf_ei|classic', '|supervised|ai_cf|classic', '|supervised|cf_dh|classic', '|supervised|ad_cf|classic', '|supervised|cf_de|classic', '|supervised|cf_fh|classic', '|supervised|cf_eh|classic', '|supervised|acf_cf|classic', '|supervised|bi_cf|classic', '|supervised|cf_cfi|classic', '|supervised|cf_cfh|classic']\nAverage final loss of the AutoVW (tuning namespaces) based on customized vw arguments: 9.606119226635231\n" + "text": [ + "Seed namespaces (singletons and interactions): ['g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "Created challengers from champion |supervised||classic|\n", + "New challenger size 37, ['|supervised|fg|classic|', '|supervised|dh|classic|', '|supervised|ef|classic|', '|supervised|ei|classic|', '|supervised|di|classic|', '|supervised|ch|classic|', '|supervised|bh|classic|', '|supervised|cf|classic|', '|supervised|ae|classic|', '|supervised|bc|classic|', '|supervised|ci|classic|', '|supervised|eg|classic|', '|supervised|ag|classic|', '|supervised|be|classic|', '|supervised|bd|classic|', '|supervised|ce|classic|', '|supervised|af|classic|', '|supervised|ad|classic|', '|supervised|ab|classic|', '|supervised|dg|classic|', '|supervised|gh|classic|', '|supervised|bg|classic|', '|supervised|fh|classic|', '|supervised|gi|classic|', '|supervised|cg|classic|', '|supervised|cd|classic|', '|supervised|ai|classic|', '|supervised|ac|classic|', '|supervised|bi|classic|', '|supervised|eh|classic|', '|supervised|fi|classic|', '|supervised|de|classic|', '|supervised|hi|classic|', '|supervised|bf|classic|', '|supervised|df|classic|', '|supervised|ah|classic|', '|supervised||classic|']\n", + "Online learning for 10000 steps...\n", + "Seed namespaces (singletons and interactions): ['df', 'g', 'a', 'h', 'b', 'c', 'i', 'd', 'e', 'f']\n", + "Created challengers from champion |supervised|df|classic|\n", + "New challenger size 43, ['|supervised|ce_df|classic|', '|supervised|df_gi|classic|', '|supervised|df_fi|classic|', '|supervised|bd_df|classic|', '|supervised|ab_df|classic|', '|supervised|bi_df|classic|', '|supervised|df_ei|classic|', '|supervised|bh_df|classic|', '|supervised|cd_df|classic|', '|supervised|df_dfg|classic|', '|supervised|def_df|classic|', '|supervised|bdf_df|classic|', '|supervised|ag_df|classic|', '|supervised|cg_df|classic|', '|supervised|df_dg|classic|', '|supervised|af_df|classic|', '|supervised|ci_df|classic|', '|supervised|df_dh|classic|', '|supervised|ah_df|classic|', '|supervised|df|classic|', '|supervised|df_di|classic|', '|supervised|ad_df|classic|', '|supervised|df_ef|classic|', '|supervised|ae_df|classic|', '|supervised|ai_df|classic|', '|supervised|be_df|classic|', '|supervised|df_eg|classic|', '|supervised|ch_df|classic|', '|supervised|ac_df|classic|', '|supervised|df_gh|classic|', '|supervised|df_fg|classic|', '|supervised|bc_df|classic|', '|supervised|df_dfh|classic|', '|supervised|df_fh|classic|', '|supervised|df_dfi|classic|', '|supervised|de_df|classic|', '|supervised|bf_df|classic|', '|supervised|bg_df|classic|', '|supervised|df_hi|classic|', '|supervised|cdf_df|classic|', '|supervised|df_eh|classic|', '|supervised|cf_df|classic|', '|supervised|adf_df|classic|']\n", + "Average final loss of the AutoVW (tuning namespaces) based on customized vw arguments: 8.828759490602918\n" + ] } ], "source": [ "''' create an AutoVW instance with ustomized VW arguments'''\n", "# parse the customized VW arguments\n", - "fixed_vw_hp_config = {'alg': 'supervised', 'loss_function': 'classic'}\n", + "fixed_vw_hp_config = {'alg': 'supervised', 'loss_function': 'classic', 'quiet': ''}\n", "search_space = fixed_vw_hp_config.copy()\n", - "search_space.update({'interactions': AutoVW.AUTO_STRING})\n", + "search_space.update({'interactions': AutoVW.AUTOMATIC,})\n", "\n", "autovw_custom = AutoVW(max_live_model_num=5, search_space=search_space) \n", "loss_list_custom = online_learning_loop(max_iter_num, vw_examples, autovw_custom)\n", @@ -390,12 +426,7 @@ "metadata": { "kernelspec": { "name": "python3", - "display_name": "Python 3", - "metadata": { - "interpreter": { - "hash": "0cfea3304185a9579d09e0953576b57c8581e46e6ebc6dfeb681bc5a511f7544" - } - } + "display_name": "Python 3.8.10 64-bit ('py38': conda)" }, "language_info": { "codemirror_mode": { @@ -407,7 +438,10 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.6.13-final" + "version": "3.8.10" + }, + "interpreter": { + "hash": "4502d015faca2560a557f35a41b6dd402f7fdfc08e843ae17a9c41947939f10c" } }, "nbformat": 4, diff --git a/setup.py b/setup.py index 462426243..75ad32fc9 100644 --- a/setup.py +++ b/setup.py @@ -56,7 +56,6 @@ setuptools.setup( "torch==1.8.1", "datasets==1.4.1", "azure-storage-blob", - "tensorflow" ], "blendsearch": [ "optuna==2.3.0" @@ -78,6 +77,7 @@ setuptools.setup( "ray[tune]>=1.2.0", "transformers", "datasets==1.4.1", + "tensorboardX<=2.2", "torch" ] },