forked from fastai/fastai
-
Notifications
You must be signed in to change notification settings - Fork 0
/
callbacks.general_sched.html
360 lines (278 loc) · 37.2 KB
/
callbacks.general_sched.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
---
title: callbacks.general_sched
keywords: fastai
sidebar: home_sidebar
summary: "Implementation of a flexible training API"
---
<div class="container" id="notebook-container">
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h2 id="TrainingPhase-and-General-scheduler">TrainingPhase and General scheduler<a class="anchor-link" href="#TrainingPhase-and-General-scheduler">¶</a></h2>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>Creates a scheduler that lets you train a model with following different <a href="/callbacks.general_sched.html#TrainingPhase"><code>TrainingPhase</code></a>.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<h2 id="TrainingPhase" class="doc_header"><code>class</code> <code>TrainingPhase</code><a href="https://github.com/fastai/fastai/blob/master/fastai/callbacks/general_sched.py#L8" class="source_link" style="float:right">[source]</a><a class="source_link" data-toggle="collapse" data-target="#TrainingPhase-pytest" style="float:right; padding-right:10px">[test]</a></h2><blockquote><p><code>TrainingPhase</code>(<strong><code>length</code></strong>:<code>int</code>)</p>
</blockquote>
<div class="collapse" id="TrainingPhase-pytest"><div class="card card-body pytest_card"><a type="button" data-toggle="collapse" data-target="#TrainingPhase-pytest" class="close" aria-label="Close"><span aria-hidden="true">×</span></a><p>No tests found for <code>TrainingPhase</code>. To contribute a test please refer to <a href="/dev/test.html">this guide</a> and <a href="https://forums.fast.ai/t/improving-expanding-functional-tests/32929">this discussion</a>.</p></div></div><p>Schedule hyper-parameters for a phase of <code>length</code> iterations.</p>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>You can then schedule any hyper-parameter you want by using the following method.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<h4 id="TrainingPhase.schedule_hp" class="doc_header"><code>schedule_hp</code><a href="https://github.com/fastai/fastai/blob/master/fastai/callbacks/general_sched.py#L13" class="source_link" style="float:right">[source]</a><a class="source_link" data-toggle="collapse" data-target="#TrainingPhase-schedule_hp-pytest" style="float:right; padding-right:10px">[test]</a></h4><blockquote><p><code>schedule_hp</code>(<strong><code>name</code></strong>, <strong><code>vals</code></strong>, <strong><code>anneal</code></strong>=<strong><em><code>None</code></em></strong>)</p>
</blockquote>
<div class="collapse" id="TrainingPhase-schedule_hp-pytest"><div class="card card-body pytest_card"><a type="button" data-toggle="collapse" data-target="#TrainingPhase-schedule_hp-pytest" class="close" aria-label="Close"><span aria-hidden="true">×</span></a><p>No tests found for <code>schedule_hp</code>. To contribute a test please refer to <a href="/dev/test.html">this guide</a> and <a href="https://forums.fast.ai/t/improving-expanding-functional-tests/32929">this discussion</a>.</p></div></div><p>Adds a schedule for <code>name</code> between <code>vals</code> using <code>anneal</code>.</p>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>The phase will make the hyper-parameter vary from the first value in <code>vals</code> to the second, following <code>anneal</code>. If an annealing function is specified but <code>vals</code> is a float, it will decay to 0. If no annealing function is specified, the default is a linear annealing for a tuple, a constant parameter if it's a float.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<div markdown="span" class="alert alert-info" role="alert"><i class="fa fa-info-circle"></i> <b>Note: </b>If you want to use discriminative values, you can pass an numpy array in `vals` (or a tuple
of them for start and stop).</div>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>The basic hyper-parameters are named:</p>
<ul>
<li>'lr' for learning rate</li>
<li>'mom' for momentum (or beta1 in Adam)</li>
<li>'beta' for the beta2 in Adam or the alpha in RMSprop</li>
<li>'wd' for weight decay</li>
</ul>
<p>You can also add any hyper-parameter that is in your optimizer (even if it's custom or a <a href="/general_optimizer.html#GeneralOptimizer"><code>GeneralOptimizer</code></a>), like 'eps' if you're using Adam.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>Let's make an example by using this to code <a href="https://arxiv.org/abs/1608.03983">SGD with warm restarts</a>.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="k">def</span> <span class="nf">fit_sgd_warm</span><span class="p">(</span><span class="n">learn</span><span class="p">,</span> <span class="n">n_cycles</span><span class="p">,</span> <span class="n">lr</span><span class="p">,</span> <span class="n">mom</span><span class="p">,</span> <span class="n">cycle_len</span><span class="p">,</span> <span class="n">cycle_mult</span><span class="p">):</span>
<span class="n">n</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">learn</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">train_dl</span><span class="p">)</span>
<span class="n">phases</span> <span class="o">=</span> <span class="p">[(</span><span class="n">TrainingPhase</span><span class="p">(</span><span class="n">n</span> <span class="o">*</span> <span class="p">(</span><span class="n">cycle_len</span> <span class="o">*</span> <span class="n">cycle_mult</span><span class="o">**</span><span class="n">i</span><span class="p">))</span>
<span class="o">.</span><span class="n">schedule_hp</span><span class="p">(</span><span class="s1">'lr'</span><span class="p">,</span> <span class="n">lr</span><span class="p">,</span> <span class="n">anneal</span><span class="o">=</span><span class="n">annealing_cos</span><span class="p">)</span>
<span class="o">.</span><span class="n">schedule_hp</span><span class="p">(</span><span class="s1">'mom'</span><span class="p">,</span> <span class="n">mom</span><span class="p">))</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">n_cycles</span><span class="p">)]</span>
<span class="n">sched</span> <span class="o">=</span> <span class="n">GeneralScheduler</span><span class="p">(</span><span class="n">learn</span><span class="p">,</span> <span class="n">phases</span><span class="p">)</span>
<span class="n">learn</span><span class="o">.</span><span class="n">callbacks</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">sched</span><span class="p">)</span>
<span class="k">if</span> <span class="n">cycle_mult</span> <span class="o">!=</span> <span class="mi">1</span><span class="p">:</span>
<span class="n">total_epochs</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">cycle_len</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="p">(</span><span class="n">cycle_mult</span><span class="p">)</span><span class="o">**</span><span class="n">n_cycles</span><span class="p">)</span><span class="o">/</span><span class="p">(</span><span class="mi">1</span><span class="o">-</span><span class="n">cycle_mult</span><span class="p">))</span>
<span class="k">else</span><span class="p">:</span> <span class="n">total_epochs</span> <span class="o">=</span> <span class="n">n_cycles</span> <span class="o">*</span> <span class="n">cycle_len</span>
<span class="n">learn</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">total_epochs</span><span class="p">)</span>
</pre></div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="n">path</span> <span class="o">=</span> <span class="n">untar_data</span><span class="p">(</span><span class="n">URLs</span><span class="o">.</span><span class="n">MNIST_SAMPLE</span><span class="p">)</span>
<span class="n">data</span> <span class="o">=</span> <span class="n">ImageDataBunch</span><span class="o">.</span><span class="n">from_folder</span><span class="p">(</span><span class="n">path</span><span class="p">)</span>
<span class="n">learn</span> <span class="o">=</span> <span class="n">Learner</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">simple_cnn</span><span class="p">((</span><span class="mi">3</span><span class="p">,</span><span class="mi">16</span><span class="p">,</span><span class="mi">16</span><span class="p">,</span><span class="mi">2</span><span class="p">)),</span> <span class="n">metrics</span><span class="o">=</span><span class="n">accuracy</span><span class="p">)</span>
<span class="n">fit_sgd_warm</span><span class="p">(</span><span class="n">learn</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mf">1e-3</span><span class="p">,</span> <span class="mf">0.9</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_html rendered_html output_subarea ">
<table border="1" class="dataframe">
<thead>
<tr style="text-align: left;">
<th>epoch</th>
<th>train_loss</th>
<th>valid_loss</th>
<th>accuracy</th>
<th>time</th>
</tr>
</thead>
<tbody>
<tr>
<td>0</td>
<td>0.162146</td>
<td>0.153532</td>
<td>0.942100</td>
<td>00:02</td>
</tr>
<tr>
<td>1</td>
<td>0.126112</td>
<td>0.117267</td>
<td>0.960255</td>
<td>00:02</td>
</tr>
<tr>
<td>2</td>
<td>0.112045</td>
<td>0.110586</td>
<td>0.962218</td>
<td>00:02</td>
</tr>
<tr>
<td>3</td>
<td>0.097603</td>
<td>0.090838</td>
<td>0.967615</td>
<td>00:02</td>
</tr>
<tr>
<td>4</td>
<td>0.086883</td>
<td>0.081375</td>
<td>0.973013</td>
<td>00:02</td>
</tr>
<tr>
<td>5</td>
<td>0.083673</td>
<td>0.076160</td>
<td>0.973994</td>
<td>00:02</td>
</tr>
<tr>
<td>6</td>
<td>0.084835</td>
<td>0.076211</td>
<td>0.973994</td>
<td>00:02</td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="input">
<div class="inner_cell">
<div class="input_area">
<div class=" highlight hl-ipython3"><pre><span></span><span class="n">learn</span><span class="o">.</span><span class="n">recorder</span><span class="o">.</span><span class="n">plot_lr</span><span class="p">()</span>
</pre></div>
</div>
</div>
</div>
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_png output_subarea ">
<img src="data:image/png;base64,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
"
>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<h2 id="GeneralScheduler" class="doc_header"><code>class</code> <code>GeneralScheduler</code><a href="https://github.com/fastai/fastai/blob/master/fastai/callbacks/general_sched.py#L18" class="source_link" style="float:right">[source]</a><a class="source_link" data-toggle="collapse" data-target="#GeneralScheduler-pytest" style="float:right; padding-right:10px">[test]</a></h2><blockquote><p><code>GeneralScheduler</code>(<strong><code>learn</code></strong>:<a href="/basic_train.html#Learner"><code>Learner</code></a>, <strong><code>phases</code></strong>:<code>Collection</code>[<a href="/callbacks.general_sched.html#TrainingPhase"><code>TrainingPhase</code></a>], <strong><code>start_epoch</code></strong>:<code>int</code>=<strong><em><code>None</code></em></strong>) :: <a href="/basic_train.html#LearnerCallback"><code>LearnerCallback</code></a></p>
</blockquote>
<div class="collapse" id="GeneralScheduler-pytest"><div class="card card-body pytest_card"><a type="button" data-toggle="collapse" data-target="#GeneralScheduler-pytest" class="close" aria-label="Close"><span aria-hidden="true">×</span></a><p>No tests found for <code>GeneralScheduler</code>. To contribute a test please refer to <a href="/dev/test.html">this guide</a> and <a href="https://forums.fast.ai/t/improving-expanding-functional-tests/32929">this discussion</a>.</p></div></div><p>Schedule multiple <a href="/callbacks.general_sched.html#TrainingPhase"><code>TrainingPhase</code></a> for a <a href="/basic_train.html#Learner"><code>Learner</code></a>.</p>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<h3 id="Callback-methods">Callback methods<a class="anchor-link" href="#Callback-methods">¶</a></h3>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>You don't call these yourself - they're called by fastai's <a href="/callback.html#Callback"><code>Callback</code></a> system automatically to enable the class's functionality.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<h4 id="GeneralScheduler.on_batch_end" class="doc_header"><code>on_batch_end</code><a href="https://github.com/fastai/fastai/blob/master/fastai/callbacks/general_sched.py#L40" class="source_link" style="float:right">[source]</a><a class="source_link" data-toggle="collapse" data-target="#GeneralScheduler-on_batch_end-pytest" style="float:right; padding-right:10px">[test]</a></h4><blockquote><p><code>on_batch_end</code>(<strong><code>train</code></strong>, <strong>**<code>kwargs</code></strong>:<code>Any</code>)</p>
</blockquote>
<div class="collapse" id="GeneralScheduler-on_batch_end-pytest"><div class="card card-body pytest_card"><a type="button" data-toggle="collapse" data-target="#GeneralScheduler-on_batch_end-pytest" class="close" aria-label="Close"><span aria-hidden="true">×</span></a><p>No tests found for <code>on_batch_end</code>. To contribute a test please refer to <a href="/dev/test.html">this guide</a> and <a href="https://forums.fast.ai/t/improving-expanding-functional-tests/32929">this discussion</a>.</p></div></div>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>Takes a step in the current phase and prepare the hyperparameters for the next batch.</p>
</div>
</div>
</div>
<div class="cell border-box-sizing code_cell rendered">
<div class="output_wrapper">
<div class="output">
<div class="output_area">
<div class="output_markdown rendered_html output_subarea ">
<h4 id="GeneralScheduler.on_train_begin" class="doc_header"><code>on_train_begin</code><a href="https://github.com/fastai/fastai/blob/master/fastai/callbacks/general_sched.py#L24" class="source_link" style="float:right">[source]</a><a class="source_link" data-toggle="collapse" data-target="#GeneralScheduler-on_train_begin-pytest" style="float:right; padding-right:10px">[test]</a></h4><blockquote><p><code>on_train_begin</code>(<strong><code>epoch</code></strong>:<code>int</code>, <strong>**<code>kwargs</code></strong>:<code>Any</code>)</p>
</blockquote>
<div class="collapse" id="GeneralScheduler-on_train_begin-pytest"><div class="card card-body pytest_card"><a type="button" data-toggle="collapse" data-target="#GeneralScheduler-on_train_begin-pytest" class="close" aria-label="Close"><span aria-hidden="true">×</span></a><p>No tests found for <code>on_train_begin</code>. To contribute a test please refer to <a href="/dev/test.html">this guide</a> and <a href="https://forums.fast.ai/t/improving-expanding-functional-tests/32929">this discussion</a>.</p></div></div>
</div>
</div>
</div>
</div>
</div>
<div class="cell border-box-sizing text_cell rendered"><div class="inner_cell">
<div class="text_cell_render border-box-sizing rendered_html">
<p>Initiates the hyperparameters to the start values of the first phase.</p>
</div>
</div>
</div>
</div>