QuickOptimizer
QuickOptimizer
#
Bases: Optimizer
QuickOptimizer implements a cost-aware Bayesian optimization. It builds upon the DyHPO algorithm, adding cost-awareness to the optimization process.
Parameters:
-
cs
(ConfigurationSpace
) –The configuration space to optimize over.
-
max_fidelity
(int
) –The maximum fidelity to optimize. Fidelity is a measure of a resource used by a configuration, such as the number of epochs.
-
perf_predictor
(PerfPredictor
, default:None
) –The performance predictor to use. If None, a new predictor is created.
-
cost_predictor
(CostPredictor
, default:None
) –The cost predictor to use. If None, a new CostPredictor is created if
cost_aware
is True. -
cost_aware
(bool
, default:False
) –Whether to use the cost predictor. Defaults to False.
-
cost_factor
(float
, default:1.0
) –A factor to control the scaling of cost values. Values must be in the range
[0.0, inf)
. A cost factor smaller than 1 compresses the cost values closer together (with 0 equalizing them), while values larger than 1 expand them. Defaults to 1.0. -
acq_fn
(str
, default:'ei'
) –The acquisition function to use. One of ["ei", "ucb", "thompson", "exploit"]. Defaults to "ei".
-
explore_factor
(float
, default:0.0
) –The exploration factor in the acquisition function. Defaults to 1.0.
-
patience
(int
, default:None
) –Determines if early stopping should be applied for a single configuration. If the score does not improve for
patience
steps, the configuration is stopped. Defaults to None. -
tol
(float
, default:0.0001
) –Tolerance for early stopping. Training stops if the score does not improve by at least
tol
forpatience
iterations (if set). Values must be in the range[0.0, inf)
. Defaults to 0.0. -
score_thresh
(float
, default:0.0
) –Threshold for early stopping. If the score is above
1 - score_thresh
, the configuration is stopped. Defaults to 0.0. -
init_random_search_steps
(int
, default:3
) –Number of configurations to evaluate randomly at the beginning of the optimization (with fidelity 1) before using predictors/acquisition function. Defaults to 10.
-
refit_init_steps
(int
, default:0
) –Number of steps (successful evaluations) before refitting the predictors. Defaults to 0.
-
refit
(bool
, default:False
) –Whether to refit the predictors with observed data. Defaults to False.
-
refit_interval
(int
, default:1
) –Interval for refitting the predictors. Defaults to 1.
-
path
(str
, default:None
) –Path to save the optimizer state. Defaults to None.
-
seed
(int
, default:None
) –Seed for reproducibility. Defaults to None.
-
verbosity
(int
, default:2
) –Verbosity level for logging. Defaults to 2.
Source code in src/qtt/optimizers/quick.py
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 |
|
ante()
#
Some operations to perform by the tuner before the optimization loop.
Here: refit the predictors with observed data.
Source code in src/qtt/optimizers/quick.py
ask()
#
Ask the optimizer for a configuration to evaluate.
Returns:
-
dict | None
–A dictionary with the configuration to evaluate.
Source code in src/qtt/optimizers/quick.py
fit(X, curve, cost)
#
Fit the predictors with the given training data.
fit_extra()
#
load(path, reset_paths=True, verbose=True)
classmethod
#
Loads the model from disk to memory.
Parameters:
-
path
(str
) –Path to the saved model, minus the file name. This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS). The model file is typically located in os.path.join(path, cls.model_file_name).
-
reset_paths
(bool
, default:True
) –Whether to reset the self.path value of the loaded model to be equal to path. It is highly recommended to keep this value as True unless accessing the original self.path value is important. If False, the actual valid path and self.path may differ, leading to strange behaviour and potential exceptions if the model needs to load any other files at a later time.
-
verbose
(bool
, default:True
) –Whether to log the location of the loaded file.
Returns:
-
model
(Optimizer
) –The loaded model object.
Source code in src/qtt/optimizers/optimizer.py
post()
#
This method is intended for the use with a tuner. It allows to perform some post-processing steps after each tell.
reset_path(path=None)
#
Reset the path of the model.
Parameters:
-
path
(str
, default:None
) –Directory location to store all outputs. If None, a new unique time-stamped directory is chosen.
Source code in src/qtt/optimizers/quick.py
save(path=None, verbose=True)
#
Saves the model to disk.
Parameters:
-
path
(str
, default:None
) –Path to the saved model, minus the file name. This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS). If None, self.path is used. The final model file is typically saved to os.path.join(path, self.model_file_name).
-
verbose
(bool
, default:True
) –Whether to log the location of the saved file.
Returns:
-
str
(str
) –Path to the saved model, minus the file name. Use this value to load the model from disk via cls.load(path), where cls is the class of the model object (e.g., model = Model.load(path)).
Source code in src/qtt/optimizers/optimizer.py
setup(n, metafeat=None)
#
Setup the optimizer for optimization.
Create the configurations to evaluate. The configurations are sampled from the configuration space. Optionally, metafeatures of the dataset can be provided.
Parameters:
-
n
(int
) –The number of configurations to create.
-
metafeat
(Mapping[str, int | float]
, default:None
) –The metafeatures of the dataset.
Source code in src/qtt/optimizers/quick.py
setup_pandas(df, metafeat=None)
#
Setup the optimizer for optimization.
Use an existing DataFrame to create the configurations to evaluate. Optionally, metafeatures of the dataset can be provided.
Parameters:
-
df
(DataFrame
) –The DataFrame with the configurations to evaluate.
-
metafeat
(Mapping[str, int | float]
, default:None
) –The metafeatures of the dataset.
Source code in src/qtt/optimizers/quick.py
tell(result)
#
Tell the result of a trial to the optimizer.
Parameters: