-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcascade_router.html
470 lines (437 loc) · 28.8 KB
/
cascade_router.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, minimum-scale=1">
<meta name="generator" content="pdoc3 0.11.1">
<title>selection.cascade_router API documentation</title>
<meta name="description" content="">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/13.0.0/sanitize.min.css" integrity="sha512-y1dtMcuvtTMJc1yPgEqF0ZjQbhnc/bFhyvIyVNb9Zk5mIGtqVaAB1Ttl28su8AvFMOY0EwRbAe+HCLqj6W7/KA==" crossorigin>
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/13.0.0/typography.min.css" integrity="sha512-Y1DYSb995BAfxobCkKepB1BqJJTPrOp3zPL74AWFugHHmmdcvO+C48WLrUOlhGMc0QG7AE3f7gmvvcrmX2fDoA==" crossorigin>
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.9.0/styles/default.min.css" crossorigin>
<style>:root{--highlight-color:#fe9}.flex{display:flex !important}body{line-height:1.5em}#content{padding:20px}#sidebar{padding:1.5em;overflow:hidden}#sidebar > *:last-child{margin-bottom:2cm}.http-server-breadcrumbs{font-size:130%;margin:0 0 15px 0}#footer{font-size:.75em;padding:5px 30px;border-top:1px solid #ddd;text-align:right}#footer p{margin:0 0 0 1em;display:inline-block}#footer p:last-child{margin-right:30px}h1,h2,h3,h4,h5{font-weight:300}h1{font-size:2.5em;line-height:1.1em}h2{font-size:1.75em;margin:2em 0 .50em 0}h3{font-size:1.4em;margin:1.6em 0 .7em 0}h4{margin:0;font-size:105%}h1:target,h2:target,h3:target,h4:target,h5:target,h6:target{background:var(--highlight-color);padding:.2em 0}a{color:#058;text-decoration:none;transition:color .2s ease-in-out}a:visited{color:#503}a:hover{color:#b62}.title code{font-weight:bold}h2[id^="header-"]{margin-top:2em}.ident{color:#900;font-weight:bold}pre code{font-size:.8em;line-height:1.4em;padding:1em;display:block}code{background:#f3f3f3;font-family:"DejaVu Sans Mono",monospace;padding:1px 4px;overflow-wrap:break-word}h1 code{background:transparent}pre{border-top:1px solid #ccc;border-bottom:1px solid #ccc;margin:1em 0}#http-server-module-list{display:flex;flex-flow:column}#http-server-module-list div{display:flex}#http-server-module-list dt{min-width:10%}#http-server-module-list p{margin-top:0}.toc ul,#index{list-style-type:none;margin:0;padding:0}#index code{background:transparent}#index h3{border-bottom:1px solid #ddd}#index ul{padding:0}#index h4{margin-top:.6em;font-weight:bold}@media (min-width:200ex){#index .two-column{column-count:2}}@media (min-width:300ex){#index .two-column{column-count:3}}dl{margin-bottom:2em}dl dl:last-child{margin-bottom:4em}dd{margin:0 0 1em 3em}#header-classes + dl > dd{margin-bottom:3em}dd dd{margin-left:2em}dd p{margin:10px 0}.name{background:#eee;font-size:.85em;padding:5px 10px;display:inline-block;min-width:40%}.name:hover{background:#e0e0e0}dt:target .name{background:var(--highlight-color)}.name > span:first-child{white-space:nowrap}.name.class > span:nth-child(2){margin-left:.4em}.inherited{color:#999;border-left:5px solid #eee;padding-left:1em}.inheritance em{font-style:normal;font-weight:bold}.desc h2{font-weight:400;font-size:1.25em}.desc h3{font-size:1em}.desc dt code{background:inherit}.source summary,.git-link-div{color:#666;text-align:right;font-weight:400;font-size:.8em;text-transform:uppercase}.source summary > *{white-space:nowrap;cursor:pointer}.git-link{color:inherit;margin-left:1em}.source pre{max-height:500px;overflow:auto;margin:0}.source pre code{font-size:12px;overflow:visible}.hlist{list-style:none}.hlist li{display:inline}.hlist li:after{content:',\2002'}.hlist li:last-child:after{content:none}.hlist .hlist{display:inline;padding-left:1em}img{max-width:100%}td{padding:0 .5em}.admonition{padding:.1em 1em;margin-bottom:1em}.admonition-title{font-weight:bold}.admonition.note,.admonition.info,.admonition.important{background:#aef}.admonition.todo,.admonition.versionadded,.admonition.tip,.admonition.hint{background:#dfd}.admonition.warning,.admonition.versionchanged,.admonition.deprecated{background:#fd4}.admonition.error,.admonition.danger,.admonition.caution{background:lightpink}</style>
<style media="screen and (min-width: 700px)">@media screen and (min-width:700px){#sidebar{width:30%;height:100vh;overflow:auto;position:sticky;top:0}#content{width:70%;max-width:100ch;padding:3em 4em;border-left:1px solid #ddd}pre code{font-size:1em}.name{font-size:1em}main{display:flex;flex-direction:row-reverse;justify-content:flex-end}.toc ul ul,#index ul ul{padding-left:1em}.toc > ul > li{margin-top:.5em}}</style>
<style media="print">@media print{#sidebar h1{page-break-before:always}.source{display:none}}@media print{*{background:transparent !important;color:#000 !important;box-shadow:none !important;text-shadow:none !important}a[href]:after{content:" (" attr(href) ")";font-size:90%}a[href][title]:after{content:none}abbr[title]:after{content:" (" attr(title) ")"}.ir a:after,a[href^="javascript:"]:after,a[href^="#"]:after{content:""}pre,blockquote{border:1px solid #999;page-break-inside:avoid}thead{display:table-header-group}tr,img{page-break-inside:avoid}img{max-width:100% !important}@page{margin:0.5cm}p,h2,h3{orphans:3;widows:3}h1,h2,h3,h4,h5,h6{page-break-after:avoid}}</style>
<script defer src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.9.0/highlight.min.js" integrity="sha512-D9gUyxqja7hBtkWpPWGt9wfbfaMGVt9gnyCvYa+jojwwPHLCzUm5i8rpk7vD7wNee9bA35eYIjobYPaQuKS1MQ==" crossorigin></script>
<script>window.addEventListener('DOMContentLoaded', () => {
hljs.configure({languages: ['bash', 'css', 'diff', 'graphql', 'ini', 'javascript', 'json', 'plaintext', 'python', 'python-repl', 'rust', 'shell', 'sql', 'typescript', 'xml', 'yaml']});
hljs.highlightAll();
})</script>
</head>
<body>
<main>
<article id="content">
<header>
<h1 class="title">Module <code>selection.cascade_router</code></h1>
</header>
<section id="section-intro">
</section>
<section>
</section>
<section>
</section>
<section>
</section>
<section>
<h2 class="section-title" id="header-classes">Classes</h2>
<dl>
<dt id="selection.cascade_router.CascadeRouter"><code class="flex name class">
<span>class <span class="ident">CascadeRouter</span></span>
<span>(</span><span>quality_computer, cost_computer, models, max_expected_cost, strategies=[<selection.lambda_strategy.ConstantStrategy object>], rounding_digits=8, greedy=False, force_order=True, max_depth=None, top_k_keep=None, set_sigma_none=False, cascade=False, do_speedup=True)</span>
</code></dt>
<dd>
<div class="desc"><p>Initializes a CascadeRouter object.</p>
<h2 id="args">Args</h2>
<dl>
<dt><strong><code>quality_computer</code></strong></dt>
<dd>The quality computer object used for computing the quality of models.</dd>
<dt><strong><code>cost_computer</code></strong></dt>
<dd>The cost computer object used for computing the cost of models.</dd>
<dt><strong><code>models</code></strong></dt>
<dd>A list of models to be considered for selection.</dd>
<dt><strong><code>max_expected_cost</code></strong></dt>
<dd>The maximum expected cost allowed for selecting models.</dd>
<dt><strong><code>strategies</code></strong></dt>
<dd>A list of hyperparameter search strategies to be used for model selection.
Default is [ConstantStrategy(10000)].</dd>
<dt><strong><code>rounding_digits</code></strong></dt>
<dd>The number of digits to round the computed values. Default is 8.</dd>
<dt><strong><code>greedy</code></strong></dt>
<dd>A boolean indicating whether to use greedy selection. Default is False.</dd>
<dt><strong><code>force_order</code></strong></dt>
<dd>A boolean indicating whether to force the execution of the models to be in the same order as the one given.
Default is True.</dd>
<dt><strong><code>max_depth</code></strong></dt>
<dd>The maximum depth allowed for supermodels in the model selection process. Default is None.</dd>
<dt><strong><code>top_k_keep</code></strong></dt>
<dd>The number of top models to keep after each step in the selection. Reduces search time.
Default is None.</dd>
<dt><strong><code>set_sigma_none</code></strong></dt>
<dd>A boolean indicating whether to set the deviations of the computed quality estimates to None.
Only used for ablation, should not be used in practice.
Default is False.</dd>
<dt><strong><code>cascade</code></strong></dt>
<dd>A boolean indicating whether to use cascading instead of cascade routing. Default is False.</dd>
<dt><strong><code>do_speedup</code></strong></dt>
<dd>A boolean indicating whether to perform speedup based on Lemma 1 in our paper.
Only used for ablation, should not be used in practice.
Default is True.</dd>
</dl></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class CascadeRouter(Algorithm):
def __init__(self, quality_computer, cost_computer,
models, max_expected_cost,
strategies=[ConstantStrategy(10000)],
rounding_digits=8, greedy=False,
force_order=True, max_depth=None,
top_k_keep=None, set_sigma_none=False,
cascade=False, do_speedup=True):
"""
Initializes a CascadeRouter object.
Args:
quality_computer: The quality computer object used for computing the quality of models.
cost_computer: The cost computer object used for computing the cost of models.
models: A list of models to be considered for selection.
max_expected_cost: The maximum expected cost allowed for selecting models.
strategies: A list of hyperparameter search strategies to be used for model selection.
Default is [ConstantStrategy(10000)].
rounding_digits: The number of digits to round the computed values. Default is 8.
greedy: A boolean indicating whether to use greedy selection. Default is False.
force_order: A boolean indicating whether to force the execution of the models to be in the same order as the one given.
Default is True.
max_depth: The maximum depth allowed for supermodels in the model selection process. Default is None.
top_k_keep: The number of top models to keep after each step in the selection. Reduces search time.
Default is None.
set_sigma_none: A boolean indicating whether to set the deviations of the computed quality estimates to None.
Only used for ablation, should not be used in practice.
Default is False.
cascade: A boolean indicating whether to use cascading instead of cascade routing. Default is False.
do_speedup: A boolean indicating whether to perform speedup based on Lemma 1 in our paper.
Only used for ablation, should not be used in practice.
Default is True.
"""
super().__init__(quality_computer, cost_computer, models,
max_expected_cost, strategies, rounding_digits)
self.lambdas = None
self.qualities = None
self.costs = None
self.gamma = None
self.greedy = greedy
self.force_order = force_order
if max_depth is not None and max_depth > len(models):
max_depth = None
self.max_depth = max_depth
self.top_k_keep = top_k_keep
self.set_sigma_none = set_sigma_none
self.cascade = cascade
self.do_speedup = do_speedup
if cascade:
self.force_order = True
def get_lambdas(self):
"""
Returns the lambdas of the cascade router.
:return: A list of lambdas.
"""
return self.lambdas
def predict(self, questions, model_answers):
qualities, sigma_qualities = self.quality_computer.predict(questions, model_answers)
costs = self.cost_computer.predict(questions, model_answers)
# sum of the first i costs is cost of ith supermodel
models = []
none_lambdas = sum([1 for lambda_ in self.lambdas if lambda_ is None])
max_depth = len(self.lambdas) - none_lambdas
if self.max_depth is not None:
max_depth = min(self.max_depth, max_depth)
for i in range(len(questions)):
step = len([j for j in model_answers[i] if j is not None])
if step >= max_depth:
models.append(None)
continue
lambda_ = self.lambdas[step]
model = self._predict_model(questions[i], qualities[i], sigma_qualities[i], costs[i],
model_answers[i], step, lambda_,
max_depth=max_depth)[0]
model = self.models[model] if model is not None else None
models.append(model)
return models
def _predict_model(self, question, qualities_question, sigma_qualities,
costs, model_answers_question, step=0,
lambda_=None, most_expensive=False,
cheapest=False, max_depth=None):
"""
Predicts the best model to run based on the given parameters.
Args:
question (any): The question to be answered.
qualities_question (list): The estimated qualities of each model for the question.
sigma_qualities (float): The deviations for the estimated qualities.
costs (list): The costs of running each model.
model_answers_question (list): The answers of each model to the question. None if the model has not been run.
step (int, optional): The current step. Defaults to 0.
lambda_ (float, optional): The lambda value. Defaults to None.
most_expensive (bool, optional): Flag indicating whether to select the most expensive model among the most optimal models.
Defaults to False.
cheapest (bool, optional): Flag indicating whether to select the cheapest model among the most optimal models.
Defaults to False.
max_depth (int, optional): The maximum depth. Defaults to None.
Returns:
tuple: A tuple containing the model to run next and the list of models to evaluate afterwards if following the same strategy.
"""
if self.max_depth is not None:
max_depth = self.max_depth if max_depth is None else min(self.max_depth, max_depth)
if max_depth is not None and step >= max_depth:
return None, []
if lambda_ is None:
lambda_ = self.lambdas[step]
if self.set_sigma_none:
sigma_qualities = None
models_already_run = [i for i in range(len(model_answers_question))
if model_answers_question[i] is not None]
models_to_evaluate = [(models_already_run, [], 0, 0)]
step = 0
best_models = dict()
while len(models_to_evaluate) > 0 and (not self.greedy or step <= 1):
next_models_to_evaluate = []
for run_models, not_run_models, quality_parent_supermodel, _ in models_to_evaluate:
all_models = run_models + not_run_models
if max_depth is not None and len(all_models) > max_depth:
continue
if len(all_models) == 0:
cost_supermodel = 0
quality_supermodel = -10 ** 8 # basically negative infinity
else:
cost_supermodel = np.sum(costs[all_models])
quality_supermodel, _ = self.quality_computer.predict_supermodels(
[question],
[all_models],
[qualities_question],
[sigma_qualities],
[model_answers_question]
)
if len(not_run_models) > 0 and self.do_speedup and not self.cascade:
cost_last_model = costs[not_run_models[-1]]
if (quality_supermodel - quality_parent_supermodel - lambda_ * cost_last_model) < 0:
continue
tradeoff = np.round(quality_supermodel - lambda_ * cost_supermodel, self.rounding_digits)
if 'all' not in best_models or tradeoff > best_models['all'][0][2]:
best_models['all'] = [(run_models, not_run_models, tradeoff, cost_supermodel)]
elif tradeoff == best_models['all'][0][2]:
best_models['all'] += [(run_models, not_run_models, tradeoff, cost_supermodel)]
models_possibilities = [i for i in range(len(model_answers_question))
if i not in all_models]
if self.force_order and len(all_models) > 0:
models_possibilities = [i for i in models_possibilities if i > all_models[-1]]
if len(not_run_models) > 0:
models_possibilities = [i for i in models_possibilities if i > not_run_models[-1]] # prevent duplicates
if self.cascade:
if len(all_models) == 0:
models_possibilities = [0]
elif all_models[-1] < len(model_answers_question) - 1:
models_possibilities = [all_models[-1] + 1]
for model in models_possibilities:
not_run_models_new = not_run_models + [model]
next_models_to_evaluate.append((run_models, not_run_models_new,
quality_supermodel, tradeoff))
step += 1
if self.top_k_keep is not None:
next_models_to_evaluate = sorted(next_models_to_evaluate, key=lambda x: x[3], reverse=True)[:self.top_k_keep]
models_to_evaluate = next_models_to_evaluate[:]
if cheapest or (not most_expensive and np.random.uniform() >= self.gamma):
best_index = np.argmin([best_models['all'][i][3] for i in range(len(best_models['all']))])
else:
best_index = np.argmax([best_models['all'][i][3] for i in range(len(best_models['all']))])
supermodel = best_models['all'][best_index]
if len(supermodel[1]) == 0:
return None, []
if not self.force_order:
index_model_to_run = np.argmin(best_models[model][2] for model in supermodel[1])
else:
index_model_to_run = 0
model_to_run = supermodel[1][index_model_to_run]
return model_to_run, supermodel[1]
def fit(self, questions, model_answers, ground_truth_qualities=None, ground_truth_costs=None):
self.quality_computer.trigger_training(True)
self.cost_computer.trigger_training(True)
self.lambdas = [0 for _ in range(len(self.models))]
if self.max_depth is not None:
self.lambdas = [0 for _ in range(self.max_depth)]
current_quality = -np.inf
for strategy in self.strategies:
lambdas, cost, quality = strategy.compute_lambdas(self.lambdas,
self._execute,
self.max_expected_cost,
(questions,
model_answers,
ground_truth_qualities,
ground_truth_costs))
if quality is not None and cost is not None and quality > current_quality and \
(cost <= self.max_expected_cost or (current_quality == -np.inf and all([lambda_ > strategy.max_lambda for lambda_ in lambdas]))):
self.lambdas = lambdas
current_quality = quality
quality_cheap,cost_cheap,quality_expensive,cost_expensive = self._execute_cheap_expensive(self.lambdas,
questions,
model_answers,
ground_truth_qualities,
ground_truth_costs)
if cost_expensive == cost_cheap:
self.gamma = 0
else:
self.gamma = (self.max_expected_cost - cost_cheap) / (cost_expensive - cost_cheap)
self.gamma = np.clip(self.gamma, 0, 1)
logger.info(f"Actual Final Lambdas: {self.lambdas}")
logger.info(f"Actual Final Cost: {(1 - self.gamma) * cost_cheap + self.gamma * cost_expensive}")
logger.info(f"Actual Final Quality: {(1 - self.gamma) * quality_cheap + self.gamma * quality_expensive}")
self.quality_computer.trigger_training(False)
self.cost_computer.trigger_training(False)
def select_answer(self, questions, model_answers):
models_selected = []
qualities, sigma_qualities = self.quality_computer.predict(questions, model_answers)
for i, quality in enumerate(qualities):
indices_with_answer = [j for j in range(len(quality)) if model_answers[i][j] is not None]
if len(indices_with_answer) == 0:
models_selected.append(None)
elif self.cascade:
models_selected.append(self.models[indices_with_answer[-1]])
else:
models_selected.append(self.models[indices_with_answer[np.argmax(quality[indices_with_answer])]])
return models_selected
def _execute_cheap_expensive(self, lambdas, questions, model_answers, ground_truth_qualities, ground_truth_costs):
output_dict_cheap = self._execute(lambdas, questions, model_answers,
cheapest=True, most_expensive=False,
ground_truth_qualities=ground_truth_qualities,
ground_truth_costs=ground_truth_costs)
quality_cheap = output_dict_cheap['quality']
cost_cheap = output_dict_cheap['cost']
output_dict_expensive = self._execute(lambdas, questions, model_answers,
cheapest=False, most_expensive=True,
ground_truth_qualities=ground_truth_qualities,
ground_truth_costs=ground_truth_costs)
quality_expensive = output_dict_expensive['quality']
cost_expensive = output_dict_expensive['cost']
return quality_cheap, cost_cheap, quality_expensive, cost_expensive
def _execute(self, lambdas, questions, model_answers,
ground_truth_qualities=None, ground_truth_costs=None,
cheapest=True, most_expensive=False):
"""
Executes the cascade router algorithm to select models based on given parameters.
Args:
lambdas (list): List of lambda values for each step in the cascade router.
questions (list): List of questions to be answered by the models.
model_answers (list): List of model answers for each question.
ground_truth_qualities (list, optional): List of ground truth qualities for each model answer. Defaults to None.
ground_truth_costs (list, optional): List of ground truth costs for each model answer. Defaults to None.
cheapest (bool, optional): Flag indicating whether to select the cheapest model. Defaults to True.
most_expensive (bool, optional): Flag indicating whether to select the most expensive model. Defaults to False.
Returns:
dict: Dictionary containing the average cost and quality of the selected models.
"""
cost = 0
quality = 0
done = [False for _ in range(len(questions))]
models_run = [[] for _ in range(len(questions))]
none_lambdas = sum([1 for lambda_ in lambdas if lambda_ is None])
max_depth = len(lambdas) - none_lambdas
for step in range(len(self.models)):
lambda_ = lambdas[step]
end_index = step
for i in range(len(questions)):
if done[i]:
continue
model_answers_sample = [model_answers[i][j] if j in models_run[i][:end_index] else None
for j in range(len(self.models))]
qualities, sigma_qualities = self.quality_computer.predict([questions[i]],
[model_answers_sample])
qualities = qualities[0]
sigma_qualities = sigma_qualities[0]
costs = self.cost_computer.predict([questions[i]], [model_answers_sample])
costs = costs[0]
model_here, future_models = self._predict_model(questions[i],
qualities, sigma_qualities,
costs, model_answers_sample,
step, lambda_,
cheapest=cheapest,
most_expensive=most_expensive,
max_depth=max_depth)
if model_here is None or step == len(self.models) - 1 or (self.max_depth is not None and step == self.max_depth - 1):
models_run_here = models_run[i] + future_models
selected_model = max(models_run_here)
if not self.cascade:
selected_model = models_run_here[np.argmax(qualities[models_run_here])]
if ground_truth_qualities is not None:
quality += ground_truth_qualities[i][selected_model]
else:
quality += qualities[selected_model]
if ground_truth_costs is not None:
cost += np.sum(ground_truth_costs[i][models_run_here])
else:
cost += np.sum(costs[models_run_here])
done[i] = True
else:
if model_here is not None:
models_run[i].append(model_here)
if all(done):
break
output_dict = {
'cost': cost / len(questions),
'quality': quality / len(questions),
}
return output_dict</code></pre>
</details>
<h3>Ancestors</h3>
<ul class="hlist">
<li><a title="selection.base_algorithm.Algorithm" href="base_algorithm.html#selection.base_algorithm.Algorithm">Algorithm</a></li>
</ul>
<h3>Methods</h3>
<dl>
<dt id="selection.cascade_router.CascadeRouter.get_lambdas"><code class="name flex">
<span>def <span class="ident">get_lambdas</span></span>(<span>self)</span>
</code></dt>
<dd>
<div class="desc"><p>Returns the lambdas of the cascade router.</p>
<p>:return: A list of lambdas.</p></div>
</dd>
</dl>
<h3>Inherited members</h3>
<ul class="hlist">
<li><code><b><a title="selection.base_algorithm.Algorithm" href="base_algorithm.html#selection.base_algorithm.Algorithm">Algorithm</a></b></code>:
<ul class="hlist">
<li><code><a title="selection.base_algorithm.Algorithm.fit" href="base_algorithm.html#selection.base_algorithm.Algorithm.fit">fit</a></code></li>
<li><code><a title="selection.base_algorithm.Algorithm.predict" href="base_algorithm.html#selection.base_algorithm.Algorithm.predict">predict</a></code></li>
<li><code><a title="selection.base_algorithm.Algorithm.select_answer" href="base_algorithm.html#selection.base_algorithm.Algorithm.select_answer">select_answer</a></code></li>
</ul>
</li>
</ul>
</dd>
</dl>
</section>
</article>
<nav id="sidebar">
<div class="toc">
<ul></ul>
</div>
<ul id="index">
<li><h3>Super-module</h3>
<ul>
<li><code><a title="selection" href="index.html">selection</a></code></li>
</ul>
</li>
<li><h3><a href="#header-classes">Classes</a></h3>
<ul>
<li>
<h4><code><a title="selection.cascade_router.CascadeRouter" href="#selection.cascade_router.CascadeRouter">CascadeRouter</a></code></h4>
<ul class="">
<li><code><a title="selection.cascade_router.CascadeRouter.get_lambdas" href="#selection.cascade_router.CascadeRouter.get_lambdas">get_lambdas</a></code></li>
</ul>
</li>
</ul>
</li>
</ul>
</nav>
</main>
<footer id="footer">
<p>Generated by <a href="https://pdoc3.github.io/pdoc" title="pdoc: Python API documentation generator"><cite>pdoc</cite> 0.11.1</a>.</p>
</footer>
</body>
</html>