-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbaseline_cascader.html
314 lines (295 loc) · 18.2 KB
/
baseline_cascader.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
<!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.baseline_cascader 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.baseline_cascader</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.baseline_cascader.BaselineCascader"><code class="flex name class">
<span>class <span class="ident">BaselineCascader</span></span>
<span>(</span><span>quality_computer, cost_computer, models, max_expected_cost, strategies=[<selection.lambda_strategy.ConstantStrategy object>])</span>
</code></dt>
<dd>
<div class="desc"><p>Initialize the BaselineCascader object. This object implements the baseline cascader,
which corresponds to the thresholding strategy for model selection in our paper.</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 model selection.</dd>
<dt><strong><code>strategies</code></strong></dt>
<dd>A list of hyperparameter search strategies to be used for model selection.
Default is [ConstantStrategy(1)].</dd>
</dl></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class BaselineCascader(Algorithm):
def __init__(self, quality_computer, cost_computer,
models, max_expected_cost,
strategies=[ConstantStrategy(1)]):
"""
Initialize the BaselineCascader object. This object implements the baseline cascader,
which corresponds to the thresholding strategy for model selection in our paper.
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 model selection.
strategies: A list of hyperparameter search strategies to be used for model selection.
Default is [ConstantStrategy(1)].
"""
super().__init__(quality_computer, cost_computer,
models, max_expected_cost,
strategies, 0)
self.lambdas = None
def get_lambdas(self):
"""
Returns the lambdas used in the baseline cascader.
Returns:
list: The lambdas used in the baseline cascader.
"""
return self.lambdas
def predict(self, questions, model_answers):
qualities, _ = self.quality_computer.predict(questions, model_answers)
models = []
for i in range(len(questions)):
step = len([answer for answer in model_answers[i] if answer is not None])
if step == 0:
models.append(self.models[0])
continue
run_next = self._predict_model(qualities[i], step)
models.append(self.models[step] if run_next > 0 else None)
return models
def select_answer(self, questions, model_answers):
models_selected = []
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)
else:
models_selected.append(self.models[np.max(indices_with_answer)])
return models_selected
def _predict_model(self, qualities_question, step=0, lambda_=None):
"""
Predicts the model based on the given qualities_question, step, and lambda_.
Parameters:
- qualities_question (list): A list of qualities for the question.
- step (int): The current step in the model selection process.
- lambda_ (float): The lambda value used for the prediction.
If not provided, it uses the lambda value corresponding to the current step.
Returns:
- int: Whether or not to continue to the next step.
"""
if step == 0:
return 1
if lambda_ is None:
lambda_ = self.lambdas[step - 1]
if qualities_question[step - 1] < 1 - lambda_:
return 1
else:
return 0
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) - 1)]
current_quality = -np.inf
qualities_per_step, costs_per_step = self.generate_step_data(questions, model_answers)
if ground_truth_qualities is None:
ground_truth_qualities = qualities_per_step[-1]
if ground_truth_costs is None:
ground_truth_costs = costs_per_step[-1]
for strategy in self.strategies:
new_lambdas, cost, quality = strategy.compute_lambdas(self.lambdas,
self._execute,
self.max_expected_cost,
(qualities_per_step,
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 new_lambdas]))):
self.lambdas = new_lambdas
current_quality = quality
elif cost is not None and cost > self.max_expected_cost:
logger.info(f"Cost {cost} is higher than maximum expected cost {self.max_expected_cost}. Stopping.")
self.quality_computer.trigger_training(False)
self.cost_computer.trigger_training(False)
def generate_step_data(self, questions, model_answers):
"""
Generates step data for the baseline cascader. This allows us to iterate quicker
in the hyperparameter optimization process.
Args:
questions (list): A list of questions.
model_answers (list): A list of model answers.
Returns:
tuple: A tuple containing two lists - qualities_per_step and costs_per_step.
- qualities_per_step (list): A list of qualities for each step.
- costs_per_step (list): A list of costs for each step.
"""
qualities_per_step = []
costs_per_step = []
for step in range(1, len(self.models)):
model_answers_step = [
list(model_answers[i][:step]) + [None for _ in range(len(model_answers[i]) - step)]
if model_answers[i] is not None else None
for i in range(len(model_answers))
]
qualities, _ = self.quality_computer.predict(questions, model_answers_step)
qualities_per_step.append(qualities)
costs = self.cost_computer.predict(questions, model_answers_step)
costs_per_step.append(costs)
return qualities_per_step, costs_per_step
def _execute(self, lambdas, qualities_step,
ground_truth_qualities, ground_truth_costs):
"""
Executes the baseline cascader algorithm on the training data.
Args:
lambdas (list): List of lambda values for each step.
qualities_step (list): List of qualities for each step.
ground_truth_qualities (list): List of ground truth qualities.
ground_truth_costs (list): List of ground truth costs.
Returns:
dict: A dictionary containing the cost and quality values.
"""
cost = 0
quality = 0
done = [False for _ in range(len(ground_truth_qualities))]
for step in range(1, len(self.models)):
step_index = step - 1
qualities = qualities_step[step_index]
lambda_ = lambdas[step_index]
for i in range(len(ground_truth_qualities)):
if done[i]:
continue
continue_here = self._predict_model(qualities[i], step, lambda_)
if continue_here == 0 or step == len(self.models) - 1:
quality += ground_truth_qualities[i][step + continue_here - 1]
cost += np.sum(ground_truth_costs[i][:step + continue_here])
done[i] = True
if all(done):
break
return {
'cost': cost / len(ground_truth_qualities),
'quality': quality / len(ground_truth_qualities),
}</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.baseline_cascader.BaselineCascader.generate_step_data"><code class="name flex">
<span>def <span class="ident">generate_step_data</span></span>(<span>self, questions, model_answers)</span>
</code></dt>
<dd>
<div class="desc"><p>Generates step data for the baseline cascader. This allows us to iterate quicker
in the hyperparameter optimization process.</p>
<h2 id="args">Args</h2>
<dl>
<dt><strong><code>questions</code></strong> : <code>list</code></dt>
<dd>A list of questions.</dd>
<dt><strong><code>model_answers</code></strong> : <code>list</code></dt>
<dd>A list of model answers.</dd>
</dl>
<h2 id="returns">Returns</h2>
<dl>
<dt><code>tuple</code></dt>
<dd>A tuple containing two lists - qualities_per_step and costs_per_step.
- qualities_per_step (list): A list of qualities for each step.
- costs_per_step (list): A list of costs for each step.</dd>
</dl></div>
</dd>
<dt id="selection.baseline_cascader.BaselineCascader.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 used in the baseline cascader.</p>
<h2 id="returns">Returns</h2>
<dl>
<dt><code>list</code></dt>
<dd>The lambdas used in the baseline cascader.</dd>
</dl></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.baseline_cascader.BaselineCascader" href="#selection.baseline_cascader.BaselineCascader">BaselineCascader</a></code></h4>
<ul class="">
<li><code><a title="selection.baseline_cascader.BaselineCascader.generate_step_data" href="#selection.baseline_cascader.BaselineCascader.generate_step_data">generate_step_data</a></code></li>
<li><code><a title="selection.baseline_cascader.BaselineCascader.get_lambdas" href="#selection.baseline_cascader.BaselineCascader.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>