├── .gitignore
├── .idea
    ├── NIPS-18.iml
    ├── markdown-navigator.xml
    ├── markdown-navigator
    │   └── profiles_settings.xml
    ├── misc.xml
    ├── modules.xml
    ├── vcs.xml
    └── workspace.xml
├── LICENSE.txt
├── Readme.md
├── coders
    ├── __init__.py
    └── vae_coding.py
├── figs
    └── train
    │   ├── grid
    │       ├── anime_samples.png
    │       ├── mnist_samples.png
    │       ├── samples.png
    │       ├── small_samples.png
    │       └── some_samples.png
    │   └── scatter
    │       └── latent.png
├── mnist
    ├── t10k-images-idx3-ubyte.gz
    ├── t10k-labels-idx1-ubyte.gz
    ├── train-images-idx3-ubyte.gz
    └── train-labels-idx1-ubyte.gz
├── models
    ├── __init__.py
    ├── generator.py
    └── vae.py
├── plots
    ├── __init__.py
    └── grid_plots.py
├── providers
    ├── __init__.py
    └── anime.py
├── sample.py
├── utils
    ├── __init__.py
    └── path.py
├── vae_train.py
├── vae_train_anime.py
└── weights
    ├── vae_anime
        ├── checkpoint
        ├── generator.data-00000-of-00001
        ├── generator.index
        └── generator.meta
    └── vae_mnist
        ├── checkpoint
        ├── generator.data-00000-of-00001
        ├── generator.index
        └── generator.meta


/.gitignore:
--------------------------------------------------------------------------------
1 | /faces
2 | 


--------------------------------------------------------------------------------
/.idea/NIPS-18.iml:
--------------------------------------------------------------------------------
 1 | <?xml version="1.0" encoding="UTF-8"?>
 2 | <module type="PYTHON_MODULE" version="4">
 3 |   <component name="NewModuleRootManager">
 4 |     <content url="file://$MODULE_DIR$" />
 5 |     <orderEntry type="inheritedJdk" />
 6 |     <orderEntry type="sourceFolder" forTests="false" />
 7 |   </component>
 8 |   <component name="TestRunnerService">
 9 |     <option name="projectConfiguration" value="py.test" />
10 |     <option name="PROJECT_TEST_RUNNER" value="py.test" />
11 |   </component>
12 | </module>


--------------------------------------------------------------------------------
/.idea/markdown-navigator.xml:
--------------------------------------------------------------------------------
 1 | <?xml version="1.0" encoding="UTF-8"?>
 2 | <project version="4">
 3 |   <component name="MarkdownProjectSettings">
 4 |     <PreviewSettings splitEditorLayout="SPLIT" splitEditorPreview="PREVIEW" useGrayscaleRendering="false" zoomFactor="1.0" maxImageWidth="0" showGitHubPageIfSynced="false" allowBrowsingInPreview="false" synchronizePreviewPosition="true" highlightPreviewType="NONE" highlightFadeOut="5" highlightOnTyping="true" synchronizeSourcePosition="true" verticallyAlignSourceAndPreviewSyncPosition="true" showSearchHighlightsInPreview="false" showSelectionInPreview="true" openRemoteLinks="true">
 5 |       <PanelProvider>
 6 |         <provider providerId="com.vladsch.idea.multimarkdown.editor.swing.html.panel" providerName="Default - Swing" />
 7 |       </PanelProvider>
 8 |     </PreviewSettings>
 9 |     <ParserSettings gitHubSyntaxChange="false">
10 |       <PegdownExtensions>
11 |         <option name="ABBREVIATIONS" value="false" />
12 |         <option name="ANCHORLINKS" value="true" />
13 |         <option name="ASIDE" value="false" />
14 |         <option name="ATXHEADERSPACE" value="true" />
15 |         <option name="AUTOLINKS" value="true" />
16 |         <option name="DEFINITIONS" value="false" />
17 |         <option name="DEFINITION_BREAK_DOUBLE_BLANK_LINE" value="false" />
18 |         <option name="FENCED_CODE_BLOCKS" value="true" />
19 |         <option name="FOOTNOTES" value="false" />
20 |         <option name="HARDWRAPS" value="false" />
21 |         <option name="HTML_DEEP_PARSER" value="false" />
22 |         <option name="INSERTED" value="false" />
23 |         <option name="QUOTES" value="false" />
24 |         <option name="RELAXEDHRULES" value="true" />
25 |         <option name="SMARTS" value="false" />
26 |         <option name="STRIKETHROUGH" value="true" />
27 |         <option name="SUBSCRIPT" value="false" />
28 |         <option name="SUPERSCRIPT" value="false" />
29 |         <option name="SUPPRESS_HTML_BLOCKS" value="false" />
30 |         <option name="SUPPRESS_INLINE_HTML" value="false" />
31 |         <option name="TABLES" value="true" />
32 |         <option name="TASKLISTITEMS" value="true" />
33 |         <option name="TOC" value="false" />
34 |         <option name="WIKILINKS" value="true" />
35 |       </PegdownExtensions>
36 |       <ParserOptions>
37 |         <option name="COMMONMARK_LISTS" value="true" />
38 |         <option name="DUMMY" value="false" />
39 |         <option name="EMOJI_SHORTCUTS" value="true" />
40 |         <option name="FLEXMARK_FRONT_MATTER" value="false" />
41 |         <option name="GFM_LOOSE_BLANK_LINE_AFTER_ITEM_PARA" value="false" />
42 |         <option name="GFM_TABLE_RENDERING" value="true" />
43 |         <option name="GITBOOK_URL_ENCODING" value="false" />
44 |         <option name="GITHUB_EMOJI_URL" value="false" />
45 |         <option name="GITHUB_LISTS" value="false" />
46 |         <option name="GITHUB_WIKI_LINKS" value="true" />
47 |         <option name="JEKYLL_FRONT_MATTER" value="false" />
48 |         <option name="SIM_TOC_BLANK_LINE_SPACER" value="true" />
49 |       </ParserOptions>
50 |     </ParserSettings>
51 |     <HtmlSettings headerTopEnabled="false" headerBottomEnabled="false" bodyTopEnabled="false" bodyBottomEnabled="false" embedUrlContent="false" addPageHeader="true" embedImages="false" embedHttpImages="false">
52 |       <GeneratorProvider>
53 |         <provider providerId="com.vladsch.idea.multimarkdown.editor.swing.html.generator" providerName="Default Swing HTML Generator" />
54 |       </GeneratorProvider>
55 |       <headerTop />
56 |       <headerBottom />
57 |       <bodyTop />
58 |       <bodyBottom />
59 |     </HtmlSettings>
60 |     <CssSettings previewScheme="UI_SCHEME" cssUri="" isCssUriEnabled="false" isCssTextEnabled="false" isDynamicPageWidth="true">
61 |       <StylesheetProvider>
62 |         <provider providerId="com.vladsch.idea.multimarkdown.editor.swing.html.css" providerName="Default Swing Stylesheet" />
63 |       </StylesheetProvider>
64 |       <ScriptProviders />
65 |       <cssText />
66 |     </CssSettings>
67 |     <HtmlExportSettings updateOnSave="false" parentDir="$ProjectFileDir$" targetDir="$ProjectFileDir$" cssDir="" scriptDir="" plainHtml="false" imageDir="" copyLinkedImages="false" imageUniquifyType="0" targetExt="" useTargetExt="false" noCssNoScripts="false" linkToExportedHtml="true" exportOnSettingsChange="true" regenerateOnProjectOpen="false" linkFormatType="HTTP_ABSOLUTE" />
68 |     <LinkMapSettings>
69 |       <textMaps />
70 |     </LinkMapSettings>
71 |   </component>
72 | </project>


--------------------------------------------------------------------------------
/.idea/markdown-navigator/profiles_settings.xml:
--------------------------------------------------------------------------------
1 | <component name="MarkdownNavigator.ProfileManager">
2 |   <settings default="" pdf-export="" />
3 | </component>


--------------------------------------------------------------------------------
/.idea/misc.xml:
--------------------------------------------------------------------------------
1 | <?xml version="1.0" encoding="UTF-8"?>
2 | <project version="4">
3 |   <component name="ProjectRootManager" version="2" project-jdk-name="Python 2.7.12 (/usr/bin/python2.7)" project-jdk-type="Python SDK" />
4 | </project>


--------------------------------------------------------------------------------
/.idea/modules.xml:
--------------------------------------------------------------------------------
1 | <?xml version="1.0" encoding="UTF-8"?>
2 | <project version="4">
3 |   <component name="ProjectModuleManager">
4 |     <modules>
5 |       <module fileurl="file://$PROJECT_DIR$/.idea/NIPS-18.iml" filepath="$PROJECT_DIR$/.idea/NIPS-18.iml" />
6 |     </modules>
7 |   </component>
8 | </project>


--------------------------------------------------------------------------------
/.idea/vcs.xml:
--------------------------------------------------------------------------------
1 | <?xml version="1.0" encoding="UTF-8"?>
2 | <project version="4">
3 |   <component name="VcsDirectoryMappings">
4 |     <mapping directory="$PROJECT_DIR$" vcs="Git" />
5 |   </component>
6 | </project>


--------------------------------------------------------------------------------
/.idea/workspace.xml:
--------------------------------------------------------------------------------
  1 | <?xml version="1.0" encoding="UTF-8"?>
  2 | <project version="4">
  3 |   <component name="ChangeListManager">
  4 |     <list default="true" id="3c04604c-ad24-4fb0-b81a-df2738d2a518" name="Default" comment="">
  5 |       <change type="MODIFICATION" beforePath="$PROJECT_DIR$/.idea/workspace.xml" afterPath="$PROJECT_DIR$/.idea/workspace.xml" />
  6 |       <change type="MODIFICATION" beforePath="$PROJECT_DIR$/models/generator.py" afterPath="$PROJECT_DIR$/models/generator.py" />
  7 |       <change type="MODIFICATION" beforePath="$PROJECT_DIR$/sample.py" afterPath="$PROJECT_DIR$/sample.py" />
  8 |     </list>
  9 |     <option name="EXCLUDED_CONVERTED_TO_IGNORED" value="true" />
 10 |     <option name="TRACKING_ENABLED" value="true" />
 11 |     <option name="SHOW_DIALOG" value="false" />
 12 |     <option name="HIGHLIGHT_CONFLICTS" value="true" />
 13 |     <option name="HIGHLIGHT_NON_ACTIVE_CHANGELIST" value="false" />
 14 |     <option name="LAST_RESOLUTION" value="IGNORE" />
 15 |   </component>
 16 |   <component name="FileEditorManager">
 17 |     <leaf SIDE_TABS_SIZE_LIMIT_KEY="300">
 18 |       <file leaf-file-name="vae.py" pinned="false" current-in-tab="false">
 19 |         <entry file="file://$PROJECT_DIR$/models/vae.py">
 20 |           <provider selected="true" editor-type-id="text-editor">
 21 |             <state relative-caret-position="486">
 22 |               <caret line="89" column="29" lean-forward="false" selection-start-line="89" selection-start-column="29" selection-end-line="89" selection-end-column="29" />
 23 |               <folding>
 24 |                 <element signature="e#0#9#0" expanded="true" />
 25 |               </folding>
 26 |             </state>
 27 |           </provider>
 28 |         </entry>
 29 |       </file>
 30 |       <file leaf-file-name="generator.py" pinned="false" current-in-tab="true">
 31 |         <entry file="file://$PROJECT_DIR$/models/generator.py">
 32 |           <provider selected="true" editor-type-id="text-editor">
 33 |             <state relative-caret-position="882">
 34 |               <caret line="49" column="16" lean-forward="false" selection-start-line="49" selection-start-column="16" selection-end-line="49" selection-end-column="16" />
 35 |               <folding>
 36 |                 <element signature="e#0#18#0" expanded="true" />
 37 |               </folding>
 38 |             </state>
 39 |           </provider>
 40 |         </entry>
 41 |       </file>
 42 |       <file leaf-file-name="sample.py" pinned="false" current-in-tab="false">
 43 |         <entry file="file://$PROJECT_DIR$/sample.py">
 44 |           <provider selected="true" editor-type-id="text-editor">
 45 |             <state relative-caret-position="216">
 46 |               <caret line="12" column="40" lean-forward="false" selection-start-line="12" selection-start-column="40" selection-end-line="12" selection-end-column="40" />
 47 |               <folding>
 48 |                 <element signature="e#0#68#0" expanded="true" />
 49 |               </folding>
 50 |             </state>
 51 |           </provider>
 52 |         </entry>
 53 |       </file>
 54 |       <file leaf-file-name="vae_coding.py" pinned="false" current-in-tab="false">
 55 |         <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
 56 |           <provider selected="true" editor-type-id="text-editor">
 57 |             <state relative-caret-position="1242">
 58 |               <caret line="72" column="15" lean-forward="false" selection-start-line="72" selection-start-column="15" selection-end-line="72" selection-end-column="15" />
 59 |               <folding>
 60 |                 <element signature="e#0#23#0" expanded="true" />
 61 |               </folding>
 62 |             </state>
 63 |           </provider>
 64 |         </entry>
 65 |       </file>
 66 |       <file leaf-file-name="Readme.md" pinned="false" current-in-tab="false">
 67 |         <entry file="file://$PROJECT_DIR$/Readme.md">
 68 |           <provider selected="true" editor-type-id="split-provider[text-editor;markdown-preview-editor]">
 69 |             <state split_layout="SPLIT">
 70 |               <first_editor relative-caret-position="0">
 71 |                 <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
 72 |                 <folding />
 73 |               </first_editor>
 74 |               <second_editor />
 75 |             </state>
 76 |           </provider>
 77 |         </entry>
 78 |       </file>
 79 |       <file leaf-file-name="vae_train_anime.py" pinned="false" current-in-tab="false">
 80 |         <entry file="file://$PROJECT_DIR$/vae_train_anime.py">
 81 |           <provider selected="true" editor-type-id="text-editor">
 82 |             <state relative-caret-position="18">
 83 |               <caret line="7" column="0" lean-forward="false" selection-start-line="7" selection-start-column="0" selection-end-line="7" selection-end-column="0" />
 84 |               <folding>
 85 |                 <element signature="e#0#68#0" expanded="true" />
 86 |               </folding>
 87 |             </state>
 88 |           </provider>
 89 |         </entry>
 90 |       </file>
 91 |       <file leaf-file-name="anime.py" pinned="false" current-in-tab="false">
 92 |         <entry file="file://$PROJECT_DIR$/providers/anime.py">
 93 |           <provider selected="true" editor-type-id="text-editor">
 94 |             <state relative-caret-position="774">
 95 |               <caret line="47" column="20" lean-forward="false" selection-start-line="47" selection-start-column="20" selection-end-line="47" selection-end-column="20" />
 96 |               <folding>
 97 |                 <element signature="e#0#21#0" expanded="true" />
 98 |               </folding>
 99 |             </state>
100 |           </provider>
101 |         </entry>
102 |       </file>
103 |       <file leaf-file-name="path.py" pinned="false" current-in-tab="false">
104 |         <entry file="file://$PROJECT_DIR$/utils/path.py">
105 |           <provider selected="true" editor-type-id="text-editor">
106 |             <state relative-caret-position="0">
107 |               <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
108 |               <folding>
109 |                 <element signature="e#0#9#0" expanded="true" />
110 |               </folding>
111 |             </state>
112 |           </provider>
113 |         </entry>
114 |       </file>
115 |       <file leaf-file-name="vae_train.py" pinned="false" current-in-tab="false">
116 |         <entry file="file://$PROJECT_DIR$/vae_train.py">
117 |           <provider selected="true" editor-type-id="text-editor">
118 |             <state relative-caret-position="0">
119 |               <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="57" selection-end-column="10" />
120 |               <folding>
121 |                 <element signature="e#0#64#0" expanded="true" />
122 |               </folding>
123 |             </state>
124 |           </provider>
125 |         </entry>
126 |       </file>
127 |       <file leaf-file-name="grid_plots.py" pinned="false" current-in-tab="false">
128 |         <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
129 |           <provider selected="true" editor-type-id="text-editor">
130 |             <state relative-caret-position="126">
131 |               <caret line="10" column="27" lean-forward="false" selection-start-line="10" selection-start-column="27" selection-end-line="10" selection-end-column="27" />
132 |               <folding>
133 |                 <element signature="e#0#18#0" expanded="true" />
134 |               </folding>
135 |             </state>
136 |           </provider>
137 |         </entry>
138 |       </file>
139 |     </leaf>
140 |   </component>
141 |   <component name="FileTemplateManagerImpl">
142 |     <option name="RECENT_TEMPLATES">
143 |       <list>
144 |         <option value="Python Script" />
145 |       </list>
146 |     </option>
147 |   </component>
148 |   <component name="Git.Settings">
149 |     <option name="RECENT_GIT_ROOT_PATH" value="$PROJECT_DIR$" />
150 |   </component>
151 |   <component name="IdeDocumentHistory">
152 |     <option name="CHANGED_PATHS">
153 |       <list>
154 |         <option value="$PROJECT_DIR$/coders/projectors.py" />
155 |         <option value="$PROJECT_DIR$/vae_train.py" />
156 |         <option value="$PROJECT_DIR$/utils/path.py" />
157 |         <option value="$PROJECT_DIR$/providers/anime.py" />
158 |         <option value="$PROJECT_DIR$/utils/plot.py" />
159 |         <option value="$PROJECT_DIR$/models/vae.py" />
160 |         <option value="$PROJECT_DIR$/var_train_anime.py" />
161 |         <option value="$PROJECT_DIR$/coders/vae_coding.py" />
162 |         <option value="$PROJECT_DIR$/plots/grid_plots.py" />
163 |         <option value="$PROJECT_DIR$/vae_train_anime.py" />
164 |         <option value="$PROJECT_DIR$/Readme.md" />
165 |         <option value="$PROJECT_DIR$/sample.py" />
166 |         <option value="$PROJECT_DIR$/models/generator.py" />
167 |       </list>
168 |     </option>
169 |   </component>
170 |   <component name="JsBuildToolGruntFileManager" detection-done="true" sorting="DEFINITION_ORDER" />
171 |   <component name="JsBuildToolPackageJson" detection-done="true" sorting="DEFINITION_ORDER" />
172 |   <component name="JsGulpfileManager">
173 |     <detection-done>true</detection-done>
174 |     <sorting>DEFINITION_ORDER</sorting>
175 |   </component>
176 |   <component name="ProjectFrameBounds" extendedState="6">
177 |     <option name="x" value="53" />
178 |     <option name="y" value="-4" />
179 |     <option name="width" value="1867" />
180 |     <option name="height" value="1084" />
181 |   </component>
182 |   <component name="ProjectLevelVcsManager" settingsEditedManually="true" />
183 |   <component name="ProjectView">
184 |     <navigator currentView="ProjectPane" proportions="" version="1">
185 |       <flattenPackages />
186 |       <showMembers />
187 |       <showModules />
188 |       <showLibraryContents />
189 |       <hideEmptyPackages />
190 |       <abbreviatePackageNames />
191 |       <autoscrollToSource />
192 |       <autoscrollFromSource />
193 |       <sortByType />
194 |       <manualOrder />
195 |       <foldersAlwaysOnTop value="true" />
196 |     </navigator>
197 |     <panes>
198 |       <pane id="Scratches" />
199 |       <pane id="ProjectPane">
200 |         <subPane>
201 |           <expand>
202 |             <path>
203 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
204 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
205 |             </path>
206 |             <path>
207 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
208 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
209 |               <item name="coders" type="462c0819:PsiDirectoryNode" />
210 |             </path>
211 |             <path>
212 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
213 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
214 |               <item name="figs" type="462c0819:PsiDirectoryNode" />
215 |             </path>
216 |             <path>
217 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
218 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
219 |               <item name="figs" type="462c0819:PsiDirectoryNode" />
220 |               <item name="train" type="462c0819:PsiDirectoryNode" />
221 |             </path>
222 |             <path>
223 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
224 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
225 |               <item name="mnist" type="462c0819:PsiDirectoryNode" />
226 |             </path>
227 |             <path>
228 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
229 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
230 |               <item name="models" type="462c0819:PsiDirectoryNode" />
231 |             </path>
232 |             <path>
233 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
234 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
235 |               <item name="plots" type="462c0819:PsiDirectoryNode" />
236 |             </path>
237 |             <path>
238 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
239 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
240 |               <item name="providers" type="462c0819:PsiDirectoryNode" />
241 |             </path>
242 |             <path>
243 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
244 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
245 |               <item name="utils" type="462c0819:PsiDirectoryNode" />
246 |             </path>
247 |             <path>
248 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
249 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
250 |               <item name="weights" type="462c0819:PsiDirectoryNode" />
251 |             </path>
252 |             <path>
253 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
254 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
255 |               <item name="weights" type="462c0819:PsiDirectoryNode" />
256 |               <item name="vae_anime" type="462c0819:PsiDirectoryNode" />
257 |             </path>
258 |             <path>
259 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="b2602c69:ProjectViewProjectNode" />
260 |               <item name="IMPLEMENTATION_Variational-Auto-Encoder" type="462c0819:PsiDirectoryNode" />
261 |               <item name="weights" type="462c0819:PsiDirectoryNode" />
262 |               <item name="vae_mnist" type="462c0819:PsiDirectoryNode" />
263 |             </path>
264 |           </expand>
265 |           <select />
266 |         </subPane>
267 |       </pane>
268 |       <pane id="Scope" />
269 |     </panes>
270 |   </component>
271 |   <component name="PropertiesComponent">
272 |     <property name="WebServerToolWindowFactoryState" value="false" />
273 |     <property name="last_opened_file_path" value="$PROJECT_DIR$" />
274 |   </component>
275 |   <component name="RunDashboard">
276 |     <option name="ruleStates">
277 |       <list>
278 |         <RuleState>
279 |           <option name="name" value="ConfigurationTypeDashboardGroupingRule" />
280 |         </RuleState>
281 |         <RuleState>
282 |           <option name="name" value="StatusDashboardGroupingRule" />
283 |         </RuleState>
284 |       </list>
285 |     </option>
286 |   </component>
287 |   <component name="ShelveChangesManager" show_recycled="false">
288 |     <option name="remove_strategy" value="false" />
289 |   </component>
290 |   <component name="TaskManager">
291 |     <task active="true" id="Default" summary="Default task">
292 |       <changelist id="3c04604c-ad24-4fb0-b81a-df2738d2a518" name="Default" comment="" />
293 |       <created>1518012270523</created>
294 |       <option name="number" value="Default" />
295 |       <option name="presentableId" value="Default" />
296 |       <updated>1518012270523</updated>
297 |     </task>
298 |     <servers />
299 |   </component>
300 |   <component name="ToolWindowManager">
301 |     <frame x="53" y="-4" width="1867" height="1084" extended-state="6" />
302 |     <layout>
303 |       <window_info id="Project" active="true" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="true" show_stripe_button="true" weight="0.17086235" sideWeight="0.5" order="0" side_tool="false" content_ui="combo" />
304 |       <window_info id="TODO" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="6" side_tool="false" content_ui="tabs" />
305 |       <window_info id="Event Log" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="7" side_tool="true" content_ui="tabs" />
306 |       <window_info id="Database" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="3" side_tool="false" content_ui="tabs" />
307 |       <window_info id="Run" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="2" side_tool="false" content_ui="tabs" />
308 |       <window_info id="Version Control" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="10" side_tool="false" content_ui="tabs" />
309 |       <window_info id="Python Console" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="8" side_tool="false" content_ui="tabs" />
310 |       <window_info id="Structure" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
311 |       <window_info id="Terminal" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="9" side_tool="false" content_ui="tabs" />
312 |       <window_info id="Debug" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.4" sideWeight="0.5" order="3" side_tool="false" content_ui="tabs" />
313 |       <window_info id="Favorites" active="false" anchor="left" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="2" side_tool="true" content_ui="tabs" />
314 |       <window_info id="Data View" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="4" side_tool="false" content_ui="tabs" />
315 |       <window_info id="Cvs" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="4" side_tool="false" content_ui="tabs" />
316 |       <window_info id="Message" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="0" side_tool="false" content_ui="tabs" />
317 |       <window_info id="Commander" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.4" sideWeight="0.5" order="0" side_tool="false" content_ui="tabs" />
318 |       <window_info id="Inspection" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.4" sideWeight="0.5" order="5" side_tool="false" content_ui="tabs" />
319 |       <window_info id="Hierarchy" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="2" side_tool="false" content_ui="combo" />
320 |       <window_info id="Find" active="false" anchor="bottom" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.33" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
321 |       <window_info id="Ant Build" active="false" anchor="right" auto_hide="false" internal_type="DOCKED" type="DOCKED" visible="false" show_stripe_button="true" weight="0.25" sideWeight="0.5" order="1" side_tool="false" content_ui="tabs" />
322 |     </layout>
323 |   </component>
324 |   <component name="TypeScriptGeneratedFilesManager">
325 |     <option name="version" value="1" />
326 |   </component>
327 |   <component name="VcsContentAnnotationSettings">
328 |     <option name="myLimit" value="2678400000" />
329 |   </component>
330 |   <component name="XDebuggerManager">
331 |     <breakpoint-manager />
332 |     <watches-manager />
333 |   </component>
334 |   <component name="editorHistoryManager">
335 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
336 |       <provider selected="true" editor-type-id="text-editor">
337 |         <state relative-caret-position="108">
338 |           <caret line="9" column="32" lean-forward="false" selection-start-line="9" selection-start-column="17" selection-end-line="9" selection-end-column="32" />
339 |           <folding>
340 |             <element signature="e#0#9#0" expanded="true" />
341 |           </folding>
342 |         </state>
343 |       </provider>
344 |     </entry>
345 |     <entry file="file://$PROJECT_DIR$/models/generator.py">
346 |       <provider selected="true" editor-type-id="text-editor">
347 |         <state relative-caret-position="432">
348 |           <caret line="26" column="34" lean-forward="false" selection-start-line="26" selection-start-column="34" selection-end-line="26" selection-end-column="34" />
349 |           <folding>
350 |             <element signature="e#0#18#0" expanded="true" />
351 |           </folding>
352 |         </state>
353 |       </provider>
354 |     </entry>
355 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
356 |       <provider selected="true" editor-type-id="text-editor">
357 |         <state relative-caret-position="1242">
358 |           <caret line="72" column="15" lean-forward="true" selection-start-line="72" selection-start-column="15" selection-end-line="72" selection-end-column="15" />
359 |           <folding>
360 |             <element signature="e#0#23#0" expanded="true" />
361 |           </folding>
362 |         </state>
363 |       </provider>
364 |     </entry>
365 |     <entry file="file://$PROJECT_DIR$/sample.py">
366 |       <provider selected="true" editor-type-id="text-editor">
367 |         <state relative-caret-position="108">
368 |           <caret line="6" column="38" lean-forward="false" selection-start-line="6" selection-start-column="38" selection-end-line="6" selection-end-column="38" />
369 |           <folding>
370 |             <element signature="e#0#68#0" expanded="true" />
371 |           </folding>
372 |         </state>
373 |       </provider>
374 |     </entry>
375 |     <entry file="file://$PROJECT_DIR$/vae_train_anime.py">
376 |       <provider selected="true" editor-type-id="text-editor">
377 |         <state relative-caret-position="18">
378 |           <caret line="7" column="0" lean-forward="false" selection-start-line="7" selection-start-column="0" selection-end-line="7" selection-end-column="0" />
379 |           <folding>
380 |             <element signature="e#0#68#0" expanded="true" />
381 |           </folding>
382 |         </state>
383 |       </provider>
384 |     </entry>
385 |     <entry file="file://$PROJECT_DIR$/providers/anime.py">
386 |       <provider selected="true" editor-type-id="text-editor">
387 |         <state relative-caret-position="774">
388 |           <caret line="47" column="20" lean-forward="true" selection-start-line="47" selection-start-column="20" selection-end-line="47" selection-end-column="20" />
389 |           <folding>
390 |             <element signature="e#0#21#0" expanded="true" />
391 |           </folding>
392 |         </state>
393 |       </provider>
394 |     </entry>
395 |     <entry file="file://$PROJECT_DIR$/utils/path.py">
396 |       <provider selected="true" editor-type-id="text-editor">
397 |         <state relative-caret-position="0">
398 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
399 |           <folding>
400 |             <element signature="e#0#9#0" expanded="true" />
401 |           </folding>
402 |         </state>
403 |       </provider>
404 |     </entry>
405 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
406 |       <provider selected="true" editor-type-id="text-editor">
407 |         <state relative-caret-position="0">
408 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="57" selection-end-column="10" />
409 |           <folding>
410 |             <element signature="e#0#64#0" expanded="true" />
411 |           </folding>
412 |         </state>
413 |       </provider>
414 |     </entry>
415 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
416 |       <provider selected="true" editor-type-id="text-editor">
417 |         <state relative-caret-position="126">
418 |           <caret line="10" column="27" lean-forward="false" selection-start-line="10" selection-start-column="27" selection-end-line="10" selection-end-column="27" />
419 |           <folding>
420 |             <element signature="e#0#18#0" expanded="true" />
421 |           </folding>
422 |         </state>
423 |       </provider>
424 |     </entry>
425 |     <entry file="file://$PROJECT_DIR$/Readme.md">
426 |       <provider selected="true" editor-type-id="split-provider[text-editor;markdown-preview-editor]">
427 |         <state split_layout="SPLIT">
428 |           <first_editor relative-caret-position="0">
429 |             <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
430 |             <folding />
431 |           </first_editor>
432 |           <second_editor />
433 |         </state>
434 |       </provider>
435 |     </entry>
436 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
437 |       <provider selected="true" editor-type-id="text-editor">
438 |         <state relative-caret-position="792">
439 |           <caret line="47" column="21" lean-forward="true" selection-start-line="47" selection-start-column="21" selection-end-line="47" selection-end-column="21" />
440 |           <folding>
441 |             <element signature="e#0#9#0" expanded="true" />
442 |           </folding>
443 |         </state>
444 |       </provider>
445 |     </entry>
446 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
447 |       <provider selected="true" editor-type-id="text-editor">
448 |         <state relative-caret-position="1314">
449 |           <caret line="74" column="0" lean-forward="true" selection-start-line="74" selection-start-column="0" selection-end-line="74" selection-end-column="0" />
450 |           <folding>
451 |             <element signature="e#0#23#0" expanded="true" />
452 |           </folding>
453 |         </state>
454 |       </provider>
455 |     </entry>
456 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
457 |       <provider selected="true" editor-type-id="text-editor">
458 |         <state relative-caret-position="918">
459 |           <caret line="57" column="10" lean-forward="true" selection-start-line="57" selection-start-column="10" selection-end-line="57" selection-end-column="10" />
460 |           <folding>
461 |             <element signature="e#0#64#0" expanded="true" />
462 |           </folding>
463 |         </state>
464 |       </provider>
465 |     </entry>
466 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
467 |       <provider selected="true" editor-type-id="text-editor">
468 |         <state relative-caret-position="936">
469 |           <caret line="55" column="0" lean-forward="false" selection-start-line="55" selection-start-column="0" selection-end-line="55" selection-end-column="22" />
470 |           <folding>
471 |             <element signature="e#0#18#0" expanded="true" />
472 |           </folding>
473 |         </state>
474 |       </provider>
475 |     </entry>
476 |     <entry file="file://$PROJECT_DIR$/Readme.md" />
477 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
478 |       <provider selected="true" editor-type-id="text-editor">
479 |         <state relative-caret-position="0">
480 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
481 |           <folding>
482 |             <element signature="e#0#9#0" expanded="true" />
483 |           </folding>
484 |         </state>
485 |       </provider>
486 |     </entry>
487 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
488 |       <provider selected="true" editor-type-id="text-editor">
489 |         <state relative-caret-position="54">
490 |           <caret line="4" column="53" lean-forward="false" selection-start-line="4" selection-start-column="53" selection-end-line="4" selection-end-column="53" />
491 |           <folding>
492 |             <element signature="e#0#23#0" expanded="true" />
493 |           </folding>
494 |         </state>
495 |       </provider>
496 |     </entry>
497 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
498 |       <provider selected="true" editor-type-id="text-editor">
499 |         <state relative-caret-position="828">
500 |           <caret line="49" column="38" lean-forward="false" selection-start-line="49" selection-start-column="30" selection-end-line="49" selection-end-column="38" />
501 |           <folding>
502 |             <element signature="e#0#18#0" expanded="true" />
503 |           </folding>
504 |         </state>
505 |       </provider>
506 |     </entry>
507 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
508 |       <provider selected="true" editor-type-id="text-editor">
509 |         <state relative-caret-position="432">
510 |           <caret line="30" column="36" lean-forward="false" selection-start-line="30" selection-start-column="36" selection-end-line="30" selection-end-column="36" />
511 |           <folding>
512 |             <element signature="e#0#64#0" expanded="true" />
513 |           </folding>
514 |         </state>
515 |       </provider>
516 |     </entry>
517 |     <entry file="file://$PROJECT_DIR$/Readme.md" />
518 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
519 |       <provider selected="true" editor-type-id="text-editor">
520 |         <state relative-caret-position="0">
521 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
522 |           <folding>
523 |             <element signature="e#0#9#0" expanded="true" />
524 |           </folding>
525 |         </state>
526 |       </provider>
527 |     </entry>
528 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
529 |       <provider selected="true" editor-type-id="text-editor">
530 |         <state relative-caret-position="54">
531 |           <caret line="4" column="53" lean-forward="false" selection-start-line="4" selection-start-column="53" selection-end-line="4" selection-end-column="53" />
532 |           <folding>
533 |             <element signature="e#0#23#0" expanded="true" />
534 |           </folding>
535 |         </state>
536 |       </provider>
537 |     </entry>
538 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
539 |       <provider selected="true" editor-type-id="text-editor">
540 |         <state relative-caret-position="828">
541 |           <caret line="49" column="38" lean-forward="false" selection-start-line="49" selection-start-column="30" selection-end-line="49" selection-end-column="38" />
542 |           <folding>
543 |             <element signature="e#0#18#0" expanded="true" />
544 |           </folding>
545 |         </state>
546 |       </provider>
547 |     </entry>
548 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
549 |       <provider selected="true" editor-type-id="text-editor">
550 |         <state relative-caret-position="432">
551 |           <caret line="30" column="36" lean-forward="false" selection-start-line="30" selection-start-column="36" selection-end-line="30" selection-end-column="36" />
552 |           <folding>
553 |             <element signature="e#0#64#0" expanded="true" />
554 |           </folding>
555 |         </state>
556 |       </provider>
557 |     </entry>
558 |     <entry file="file://$PROJECT_DIR$/Readme.md" />
559 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
560 |       <provider selected="true" editor-type-id="text-editor">
561 |         <state relative-caret-position="0">
562 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
563 |           <folding>
564 |             <element signature="e#0#9#0" expanded="true" />
565 |           </folding>
566 |         </state>
567 |       </provider>
568 |     </entry>
569 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
570 |       <provider selected="true" editor-type-id="text-editor">
571 |         <state relative-caret-position="54">
572 |           <caret line="4" column="53" lean-forward="false" selection-start-line="4" selection-start-column="53" selection-end-line="4" selection-end-column="53" />
573 |           <folding>
574 |             <element signature="e#0#23#0" expanded="true" />
575 |           </folding>
576 |         </state>
577 |       </provider>
578 |     </entry>
579 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
580 |       <provider selected="true" editor-type-id="text-editor">
581 |         <state relative-caret-position="432">
582 |           <caret line="30" column="36" lean-forward="false" selection-start-line="30" selection-start-column="36" selection-end-line="30" selection-end-column="36" />
583 |           <folding>
584 |             <element signature="e#0#64#0" expanded="true" />
585 |           </folding>
586 |         </state>
587 |       </provider>
588 |     </entry>
589 |     <entry file="file://$PROJECT_DIR$/Readme.md" />
590 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
591 |       <provider selected="true" editor-type-id="text-editor">
592 |         <state relative-caret-position="1980">
593 |           <caret line="112" column="37" lean-forward="false" selection-start-line="112" selection-start-column="37" selection-end-line="112" selection-end-column="37" />
594 |           <folding>
595 |             <element signature="e#0#9#0" expanded="true" />
596 |           </folding>
597 |         </state>
598 |       </provider>
599 |     </entry>
600 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
601 |       <provider selected="true" editor-type-id="text-editor">
602 |         <state relative-caret-position="54">
603 |           <caret line="4" column="53" lean-forward="false" selection-start-line="4" selection-start-column="53" selection-end-line="4" selection-end-column="53" />
604 |           <folding>
605 |             <element signature="e#0#23#0" expanded="true" />
606 |           </folding>
607 |         </state>
608 |       </provider>
609 |     </entry>
610 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
611 |       <provider selected="true" editor-type-id="text-editor">
612 |         <state relative-caret-position="378">
613 |           <caret line="27" column="63" lean-forward="false" selection-start-line="27" selection-start-column="63" selection-end-line="27" selection-end-column="63" />
614 |           <folding>
615 |             <element signature="e#0#64#0" expanded="true" />
616 |           </folding>
617 |         </state>
618 |       </provider>
619 |     </entry>
620 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
621 |       <provider selected="true" editor-type-id="text-editor">
622 |         <state relative-caret-position="414">
623 |           <caret line="26" column="22" lean-forward="false" selection-start-line="26" selection-start-column="22" selection-end-line="26" selection-end-column="22" />
624 |           <folding>
625 |             <element signature="e#0#18#0" expanded="true" />
626 |           </folding>
627 |         </state>
628 |       </provider>
629 |     </entry>
630 |     <entry file="file://$PROJECT_DIR$/Readme.md" />
631 |     <entry file="file://$PROJECT_DIR$/coders/projectors.py" />
632 |     <entry file="file://$PROJECT_DIR$/utils/plot.py" />
633 |     <entry file="file://$PROJECT_DIR$/utils/path.py">
634 |       <provider selected="true" editor-type-id="text-editor">
635 |         <state relative-caret-position="0">
636 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
637 |           <folding>
638 |             <element signature="e#0#9#0" expanded="true" />
639 |           </folding>
640 |         </state>
641 |       </provider>
642 |     </entry>
643 |     <entry file="file://$PROJECT_DIR$/providers/anime.py">
644 |       <provider selected="true" editor-type-id="text-editor">
645 |         <state relative-caret-position="774">
646 |           <caret line="47" column="20" lean-forward="false" selection-start-line="47" selection-start-column="20" selection-end-line="47" selection-end-column="20" />
647 |           <folding>
648 |             <element signature="e#0#21#0" expanded="true" />
649 |           </folding>
650 |         </state>
651 |       </provider>
652 |     </entry>
653 |     <entry file="file://$PROJECT_DIR$/vae_train.py">
654 |       <provider selected="true" editor-type-id="text-editor">
655 |         <state relative-caret-position="0">
656 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="57" selection-end-column="10" />
657 |           <folding>
658 |             <element signature="e#0#64#0" expanded="true" />
659 |           </folding>
660 |         </state>
661 |       </provider>
662 |     </entry>
663 |     <entry file="file://$PROJECT_DIR$/plots/grid_plots.py">
664 |       <provider selected="true" editor-type-id="text-editor">
665 |         <state relative-caret-position="126">
666 |           <caret line="10" column="27" lean-forward="false" selection-start-line="10" selection-start-column="27" selection-end-line="10" selection-end-column="27" />
667 |           <folding>
668 |             <element signature="e#0#18#0" expanded="true" />
669 |           </folding>
670 |         </state>
671 |       </provider>
672 |     </entry>
673 |     <entry file="file://$PROJECT_DIR$/coders/vae_coding.py">
674 |       <provider selected="true" editor-type-id="text-editor">
675 |         <state relative-caret-position="1242">
676 |           <caret line="72" column="15" lean-forward="false" selection-start-line="72" selection-start-column="15" selection-end-line="72" selection-end-column="15" />
677 |           <folding>
678 |             <element signature="e#0#23#0" expanded="true" />
679 |           </folding>
680 |         </state>
681 |       </provider>
682 |     </entry>
683 |     <entry file="file://$PROJECT_DIR$/vae_train_anime.py">
684 |       <provider selected="true" editor-type-id="text-editor">
685 |         <state relative-caret-position="18">
686 |           <caret line="7" column="0" lean-forward="false" selection-start-line="7" selection-start-column="0" selection-end-line="7" selection-end-column="0" />
687 |           <folding>
688 |             <element signature="e#0#68#0" expanded="true" />
689 |           </folding>
690 |         </state>
691 |       </provider>
692 |     </entry>
693 |     <entry file="file://$PROJECT_DIR$/Readme.md">
694 |       <provider editor-type-id="text-editor">
695 |         <state relative-caret-position="0">
696 |           <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
697 |           <folding />
698 |         </state>
699 |       </provider>
700 |       <provider selected="true" editor-type-id="split-provider[text-editor;markdown-preview-editor]">
701 |         <state split_layout="SPLIT">
702 |           <first_editor relative-caret-position="0">
703 |             <caret line="0" column="0" lean-forward="false" selection-start-line="0" selection-start-column="0" selection-end-line="0" selection-end-column="0" />
704 |             <folding />
705 |           </first_editor>
706 |           <second_editor />
707 |         </state>
708 |       </provider>
709 |     </entry>
710 |     <entry file="file://$PROJECT_DIR$/models/vae.py">
711 |       <provider selected="true" editor-type-id="text-editor">
712 |         <state relative-caret-position="486">
713 |           <caret line="89" column="29" lean-forward="false" selection-start-line="89" selection-start-column="29" selection-end-line="89" selection-end-column="29" />
714 |           <folding>
715 |             <element signature="e#0#9#0" expanded="true" />
716 |           </folding>
717 |         </state>
718 |       </provider>
719 |     </entry>
720 |     <entry file="file://$PROJECT_DIR$/sample.py">
721 |       <provider selected="true" editor-type-id="text-editor">
722 |         <state relative-caret-position="216">
723 |           <caret line="12" column="40" lean-forward="false" selection-start-line="12" selection-start-column="40" selection-end-line="12" selection-end-column="40" />
724 |           <folding>
725 |             <element signature="e#0#68#0" expanded="true" />
726 |           </folding>
727 |         </state>
728 |       </provider>
729 |     </entry>
730 |     <entry file="file://$PROJECT_DIR$/models/generator.py">
731 |       <provider selected="true" editor-type-id="text-editor">
732 |         <state relative-caret-position="882">
733 |           <caret line="49" column="16" lean-forward="false" selection-start-line="49" selection-start-column="16" selection-end-line="49" selection-end-column="16" />
734 |           <folding>
735 |             <element signature="e#0#18#0" expanded="true" />
736 |           </folding>
737 |         </state>
738 |       </provider>
739 |     </entry>
740 |   </component>
741 | </project>


--------------------------------------------------------------------------------
/LICENSE.txt:
--------------------------------------------------------------------------------
 1 | MIT License
 2 | 
 3 | Copyright (c) [2018] [Ga Wu] [University of Toronto]
 4 | 
 5 | Permission is hereby granted, free of charge, to any person obtaining a copy
 6 | of this software and associated documentation files (the "Software"), to deal
 7 | in the Software without restriction, including without limitation the rights
 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 9 | copies of the Software, and to permit persons to whom the Software is
10 | furnished to do so, subject to the following conditions:
11 | 
12 | The above copyright notice and this permission notice shall be included in all
13 | copies or substantial portions of the Software.
14 | 
15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 | SOFTWARE.
22 | 


--------------------------------------------------------------------------------
/Readme.md:
--------------------------------------------------------------------------------
 1 | Variational Autoencoder
 2 | ===
 3 | This is a enhanced implementation of Variational Autoencoder.
 4 | Both fully connected and convolutional encoder/decoder are built in this model.
 5 | Please star if you like this implementation.
 6 | 
 7 | ## Use
 8 | ```python
 9 | $python vae_train_amine.py # for training
10 | $python sample.py # for sampling
11 | ```
12 | 
13 | ## Update
14 | 1. Removed standard derivation learning on Gaussian observation decoder.
15 | 2. Set the standard derivation of observation to hyper-parameter.
16 | 3. Add deconvolution CNN support for the Anime dataset.
17 | 4. Remove Anime dataset itself to avoid legal issues.
18 | 
19 | ## Pre-Trained Models
20 | There are two pretrained models
21 | 1. Anime
22 | 2. MNIST
23 | 
24 | The weights of pretrained models are locaded in weights folder
25 | 
26 | ## Samples
27 | 
28 | ### ANIME
29 | ![](https://github.com/wuga214/Variational-Auto-Encoder/blob/master/figs/train/grid/anime_samples.png)
30 | 
31 | ### MNIST
32 | ![](https://github.com/wuga214/Variational-Auto-Encoder/blob/master/figs/train/grid/mnist_samples.png)
33 | 
34 | 
35 | ### Latent Space Distribution
36 | ![](https://github.com/wuga214/Variational-Auto-Encoder/blob/master/figs/train/scatter/latent.png)
37 | 


--------------------------------------------------------------------------------
/coders/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/coders/__init__.py


--------------------------------------------------------------------------------
/coders/vae_coding.py:
--------------------------------------------------------------------------------
  1 | import tensorflow as tf
  2 | import tensorlayer as tl
  3 | from tensorlayer.layers import *
  4 | from tensorflow.contrib import layers
  5 | 
  6 | 
  7 | def fc_mnist_encoder(x, latent_dim, activation=None):
  8 |     e = layers.fully_connected(x, 500, scope='fc-01')
  9 |     e = layers.fully_connected(e, 500, scope='fc-02')
 10 |     e = layers.fully_connected(e, 200, scope='fc-03')
 11 |     output = layers.fully_connected(e, 2 * latent_dim, activation_fn=activation,
 12 |                                     scope='fc-final')
 13 | 
 14 |     return output
 15 | 
 16 | 
 17 | def fc_mnist_decoder(z, observation_dim, activation=tf.sigmoid):
 18 |     x = layers.fully_connected(z, 200, scope='fc-01')
 19 |     x = layers.fully_connected(x, 500, scope='fc-02')
 20 |     x = layers.fully_connected(x, 500, scope='fc-03')
 21 |     output = layers.fully_connected(x, observation_dim, activation_fn=activation,
 22 |                                     scope='fc-final')
 23 |     return output
 24 | 
 25 | def conv_anime_encoder(x, latent_dim, activation=None):
 26 |     is_train = True
 27 |     z_dim = latent_dim  # 512
 28 |     ef_dim = 64  # encoder filter number
 29 | 
 30 |     w_init = tf.random_normal_initializer(stddev=0.02)
 31 |     gamma_init = tf.random_normal_initializer(1., 0.02)
 32 | 
 33 |     input = tf.reshape(x, [-1, 64, 64, 3])
 34 |     net_in = InputLayer(input, name='en/in')  # (b_size,64,64,3)
 35 |     net_h0 = Conv2d(net_in, ef_dim, (5, 5), (2, 2), act=None,
 36 |                     padding='SAME', W_init=w_init, name='en/h0/conv2d')
 37 |     net_h0 = BatchNormLayer(net_h0, act=tf.nn.relu,
 38 |                             is_train=is_train, gamma_init=gamma_init, name='en/h0/batch_norm')
 39 | 
 40 |     net_h1 = Conv2d(net_h0, ef_dim * 2, (5, 5), (2, 2), act=None,
 41 |                     padding='SAME', W_init=w_init, name='en/h1/conv2d')
 42 |     net_h1 = BatchNormLayer(net_h1, act=tf.nn.relu,
 43 |                             is_train=is_train, gamma_init=gamma_init, name='en/h1/batch_norm')
 44 | 
 45 |     net_h2 = Conv2d(net_h1, ef_dim * 4, (5, 5), (2, 2), act=None,
 46 |                     padding='SAME', W_init=w_init, name='en/h2/conv2d')
 47 |     net_h2 = BatchNormLayer(net_h2, act=tf.nn.relu,
 48 |                             is_train=is_train, gamma_init=gamma_init, name='en/h2/batch_norm')
 49 | 
 50 |     net_h3 = Conv2d(net_h2, ef_dim * 8, (5, 5), (2, 2), act=None,
 51 |                     padding='SAME', W_init=w_init, name='en/h3/conv2d')
 52 |     net_h3 = BatchNormLayer(net_h3, act=tf.nn.relu,
 53 |                             is_train=is_train, gamma_init=gamma_init, name='en/h3/batch_norm')
 54 | 
 55 | 
 56 |     net_h4 = FlattenLayer(net_h3, name='en/h4/flatten')
 57 | 
 58 |     net_out = DenseLayer(net_h4, n_units=z_dim*2, act=tf.identity,
 59 |                          W_init=w_init, name='en/h3/lin_sigmoid')
 60 |     net_out = BatchNormLayer(net_out, act=tf.identity,
 61 |                              is_train=is_train, gamma_init=gamma_init, name='en/out1/batch_norm')
 62 | 
 63 |     output = net_out.outputs
 64 | 
 65 |     return output
 66 | 
 67 | 
 68 | def conv_anime_decoder(z, observation_dim, activation=tf.tanh):
 69 | 
 70 |     is_train = True
 71 |     image_size = 64
 72 |     s2, s4, s8, s16 = int(image_size / 2), int(image_size / 4), int(image_size / 8), int(image_size / 16)  # 32,16,8,4
 73 |     gf_dim = 64
 74 |     c_dim = 3
 75 | 
 76 |     w_init = tf.random_normal_initializer(stddev=0.02)
 77 |     gamma_init = tf.random_normal_initializer(1., 0.02)
 78 | 
 79 |     net_in = InputLayer(z, name='g/in')
 80 |     net_h0 = DenseLayer(net_in, n_units=gf_dim * 4 * s8 * s8, W_init=w_init,
 81 |                         act=tf.identity, name='g/h0/lin')
 82 |     # net_h0.outputs._shape = (b_size,256*8*8)
 83 |     net_h0 = ReshapeLayer(net_h0, shape=[-1, s8, s8, gf_dim * 4], name='g/h0/reshape')
 84 |     # net_h0.outputs._shape = (b_size,8,8,256)
 85 |     net_h0 = BatchNormLayer(net_h0, act=tf.nn.relu, is_train=is_train,
 86 |                             gamma_init=gamma_init, name='g/h0/batch_norm')
 87 | 
 88 |     # upsampling
 89 |     net_h1 = DeConv2d(net_h0, gf_dim * 4, (5, 5), out_size=(s4, s4), strides=(2, 2),
 90 |                       padding='SAME', act=None, W_init=w_init, name='g/h1/decon2d')
 91 |     net_h1 = BatchNormLayer(net_h1, act=tf.nn.relu, is_train=is_train,
 92 |                             gamma_init=gamma_init, name='g/h1/batch_norm')
 93 |     # net_h1.outputs._shape = (b_size,16,16,256)
 94 | 
 95 |     net_h2 = DeConv2d(net_h1, gf_dim * 2, (5, 5), out_size=(s2, s2), strides=(2, 2),
 96 |                       padding='SAME', act=None, W_init=w_init, name='g/h2/decon2d')
 97 |     net_h2 = BatchNormLayer(net_h2, act=tf.nn.relu, is_train=is_train,
 98 |                             gamma_init=gamma_init, name='g/h2/batch_norm')
 99 |     # net_h2.outputs._shape = (b_size,32,32,128)
100 | 
101 |     net_h3 = DeConv2d(net_h2, gf_dim // 2, (5, 5), out_size=(image_size, image_size), strides=(2, 2),
102 |                       padding='SAME', act=None, W_init=w_init, name='g/h3/decon2d')
103 |     net_h3 = BatchNormLayer(net_h3, act=tf.nn.relu, is_train=is_train,
104 |                             gamma_init=gamma_init, name='g/h3/batch_norm')
105 |     # net_h3.outputs._shape = (b_size,64,64,32)
106 | 
107 |     # no BN on last deconv
108 |     net_h4 = DeConv2d(net_h3, c_dim, (5, 5), out_size=(image_size, image_size), strides=(1, 1),
109 |                       padding='SAME', act=None, W_init=w_init, name='g/h4/decon2d')
110 |     # net_h4.outputs._shape = (b_size,64,64,3)
111 | 
112 |     logits = net_h4.outputs
113 |     net_h4.outputs = tf.nn.tanh(net_h4.outputs)
114 | 
115 |     output = layers.flatten(net_h4.outputs)
116 | 
117 |     return output
118 | 


--------------------------------------------------------------------------------
/figs/train/grid/anime_samples.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/grid/anime_samples.png


--------------------------------------------------------------------------------
/figs/train/grid/mnist_samples.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/grid/mnist_samples.png


--------------------------------------------------------------------------------
/figs/train/grid/samples.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/grid/samples.png


--------------------------------------------------------------------------------
/figs/train/grid/small_samples.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/grid/small_samples.png


--------------------------------------------------------------------------------
/figs/train/grid/some_samples.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/grid/some_samples.png


--------------------------------------------------------------------------------
/figs/train/scatter/latent.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/figs/train/scatter/latent.png


--------------------------------------------------------------------------------
/mnist/t10k-images-idx3-ubyte.gz:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/mnist/t10k-images-idx3-ubyte.gz


--------------------------------------------------------------------------------
/mnist/t10k-labels-idx1-ubyte.gz:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/mnist/t10k-labels-idx1-ubyte.gz


--------------------------------------------------------------------------------
/mnist/train-images-idx3-ubyte.gz:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/mnist/train-images-idx3-ubyte.gz


--------------------------------------------------------------------------------
/mnist/train-labels-idx1-ubyte.gz:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/mnist/train-labels-idx1-ubyte.gz


--------------------------------------------------------------------------------
/models/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/models/__init__.py


--------------------------------------------------------------------------------
/models/generator.py:
--------------------------------------------------------------------------------
 1 | import numpy as np
 2 | import tensorflow as tf
 3 | from tensorflow.contrib.distributions import Normal
 4 | 
 5 | 
 6 | class GENERATOR(object):
 7 |     def __init__(self, latent_dim, observation_dim, generator,
 8 |                  obs_distrib="Bernoulli",
 9 |                  obs_std=0.01,
10 |                  ):
11 |         """
12 | 
13 |         """
14 |         self._latent_dim = latent_dim
15 |         self._observation_dim = observation_dim
16 |         self._generator = generator
17 |         self._obs_distrib = obs_distrib
18 |         self._obs_std = obs_std
19 |         self._p_distribution = self._multivariate_normal()
20 |         self._build_graph()
21 | 
22 |     def _multivariate_normal(self):
23 |         return Normal([0.] * self._latent_dim, [1.] * self._latent_dim)
24 | 
25 |     def _build_graph(self):
26 | 
27 |         with tf.variable_scope('in'):
28 |             # placeholder for the input noise
29 |             self.candid = tf.placeholder(tf.float32, shape=[None, self._latent_dim], name='candidate')
30 | 
31 |             # decode batch
32 |             with tf.variable_scope('generator'):
33 |                 self.generated = self._generator(self.candid, self._latent_dim)
34 | 
35 |             self._sesh = tf.Session()
36 |             init = tf.global_variables_initializer()
37 |             self._sesh.run(init)
38 | 
39 |     def load_pretrained(self, path):
40 |         generator_variables = []
41 |         for v in tf.trainable_variables():
42 |             if "generator" in v.name:
43 |                 generator_variables.append(v)
44 |         saver = tf.train.Saver(generator_variables)
45 |         saver.restore(self._sesh, path)
46 | 
47 |     def e2x(self, noise):
48 |         x = self._sesh.run(self.generated,
49 |                            feed_dict={self.candid: noise})
50 |         return x


--------------------------------------------------------------------------------
/models/vae.py:
--------------------------------------------------------------------------------
  1 | import re
  2 | import numpy as np
  3 | import tensorflow as tf
  4 | from tensorflow.contrib.distributions import Bernoulli
  5 | 
  6 | 
  7 | class VAE(object):
  8 | 
  9 |     def __init__(self, latent_dim, batch_size, encoder, decoder,
 10 |                  observation_dim=784,
 11 |                  learning_rate=1e-4,
 12 |                  optimizer=tf.train.RMSPropOptimizer,
 13 |                  observation_distribution="Bernoulli", # or Gaussian
 14 |                  observation_std=0.01):
 15 | 
 16 |         self._latent_dim = latent_dim
 17 |         self._batch_size = batch_size
 18 |         self._encode = encoder
 19 |         self._decode = decoder
 20 |         self._observation_dim = observation_dim
 21 |         self._learning_rate = learning_rate
 22 |         self._optimizer = optimizer
 23 |         self._observation_distribution = observation_distribution
 24 |         self._observation_std = observation_std
 25 |         self._build_graph()
 26 | 
 27 |     def _build_graph(self):
 28 | 
 29 |         with tf.variable_scope('vae'):
 30 |             self.x = tf.placeholder(tf.float32, shape=[None, self._observation_dim])
 31 | 
 32 |             with tf.variable_scope('encoder'):
 33 |                 encoded = self._encode(self.x, self._latent_dim)
 34 | 
 35 |             with tf.variable_scope('latent'):
 36 |                 self.mean = encoded[:, :self._latent_dim]
 37 |                 logvar = encoded[:, self._latent_dim:]
 38 |                 stddev = tf.sqrt(tf.exp(logvar))
 39 |                 epsilon = tf.random_normal([self._batch_size, self._latent_dim])
 40 |                 self.z = self.mean + stddev * epsilon
 41 | 
 42 |             with tf.variable_scope('decoder'):
 43 |                 decoded = self._decode(self.z, self._observation_dim)
 44 |                 self.obs_mean = decoded
 45 |                 if self._observation_distribution == 'Gaussian':
 46 |                     obs_epsilon = tf.random_normal([self._batch_size, self._observation_dim])
 47 |                     self.sample = self.obs_mean + self._observation_std * obs_epsilon
 48 |                 else:
 49 |                     self.sample = Bernoulli(probs=self.obs_mean).sample()
 50 | 
 51 | 
 52 |             with tf.variable_scope('loss'):
 53 |                 with tf.variable_scope('kl-divergence'):
 54 |                     kl = self._kl_diagnormal_stdnormal(self.mean, logvar)
 55 | 
 56 |                 if self._observation_distribution == 'Gaussian':
 57 |                     with tf.variable_scope('gaussian'):
 58 |                         obj = self._gaussian_log_likelihood(self.x, self.obs_mean, self._observation_std)
 59 |                 else:
 60 |                     with tf.variable_scope('bernoulli'):
 61 |                         obj = self._bernoulli_log_likelihood(self.x, self.obs_mean)
 62 | 
 63 |                 self._loss = (kl + obj) / self._batch_size
 64 | 
 65 |             with tf.variable_scope('optimizer'):
 66 |                 optimizer = tf.train.RMSPropOptimizer(learning_rate=self._learning_rate)
 67 |             with tf.variable_scope('training-step'):
 68 |                 self._train = optimizer.minimize(self._loss)
 69 | 
 70 |             self._sesh = tf.Session()
 71 |             init = tf.global_variables_initializer()
 72 |             self._sesh.run(init)
 73 | 
 74 |     @staticmethod
 75 |     def _kl_diagnormal_stdnormal(mu, log_var):
 76 | 
 77 |         var = tf.exp(log_var)
 78 |         kl = 0.5 * tf.reduce_sum(tf.square(mu) + var - 1. - log_var)
 79 |         return kl
 80 | 
 81 |     @staticmethod
 82 |     def _gaussian_log_likelihood(targets, mean, std):
 83 |         se = 0.5 * tf.reduce_sum(tf.square(targets - mean)) / (2*tf.square(std)) + tf.log(std)
 84 |         return se
 85 | 
 86 |     @staticmethod
 87 |     def _bernoulli_log_likelihood(targets, outputs, eps=1e-8):
 88 | 
 89 |         log_like = -tf.reduce_sum(targets * tf.log(outputs + eps)
 90 |                                   + (1. - targets) * tf.log((1. - outputs) + eps))
 91 |         return log_like
 92 | 
 93 |     def update(self, x):
 94 |         _, loss = self._sesh.run([self._train, self._loss], feed_dict={self.x: x})
 95 |         return loss
 96 | 
 97 |     def x2z(self, x):
 98 | 
 99 |         mean = self._sesh.run([self.mean], feed_dict={self.x: x})
100 | 
101 |         return np.asarray(mean).reshape(-1, self._latent_dim)
102 | 
103 |     def z2x(self, z):
104 | 
105 |         x = self._sesh.run([self.obs_mean], feed_dict={self.z: z})
106 |         return x
107 | 
108 |     def sample(self, z):
109 | 
110 |         x = self._sesh.run([self.sample], feed_dict={self.z: z})
111 | 
112 |         return x
113 | 
114 |     def save_generator(self, path, prefix="in/generator"):
115 |         variables = tf.trainable_variables()
116 |         var_dict = {}
117 |         for v in variables:
118 |             if "decoder" in v.name:
119 |                 name = prefix+re.sub("vae/decoder", "", v.name)
120 |                 name = re.sub(":0", "", name)
121 |                 var_dict[name] = v
122 |         for k, v in var_dict.items():
123 |             print(k)
124 |             print(v)
125 |         saver = tf.train.Saver(var_dict)
126 |         saver.save(self._sesh, path)


--------------------------------------------------------------------------------
/plots/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/plots/__init__.py


--------------------------------------------------------------------------------
/plots/grid_plots.py:
--------------------------------------------------------------------------------
 1 | import numpy as np
 2 | import matplotlib.pyplot as plt
 3 | from mpl_toolkits.axes_grid1 import ImageGrid
 4 | import seaborn as sns
 5 | sns.axes_style("white")
 6 | 
 7 | 
 8 | def show_samples(images, row, col, image_shape, name="Unknown", save=True, shift=False):
 9 |     num_images = row*col
10 |     if shift:
11 |         images = (images+1.)/2.
12 |     fig = plt.figure(figsize=(col, row))
13 |     grid = ImageGrid(fig, 111,
14 |                      nrows_ncols=(row, col),
15 |                      axes_pad=0.)
16 |     for i in xrange(num_images):
17 |         im = images[i].reshape(image_shape)
18 |         axis = grid[i]
19 |         axis.axis('off')
20 |         axis.imshow(im)
21 |     plt.axis('off')
22 |     plt.tight_layout()
23 |     if save:
24 |         fig.savefig('figs/train/grid/'+name+'.png', bbox_inches="tight", pad_inches=0, format='png')
25 |     else:
26 |         plt.show()
27 | 
28 | 
29 | #From some github code
30 | def show_latent_scatter(vae, data, name="latent"):
31 |     n_test = 5000
32 |     batch_size = 100
33 |     zs = np.zeros((n_test, 2), dtype=np.float32)
34 |     labels = np.zeros(n_test)
35 |     for i in range(int(n_test / batch_size)):
36 |         x, y = data.test.next_batch(batch_size)
37 |         labels[(100 * i):(100 * (i + 1))] = y
38 |         z = vae.x2z(x)
39 |         zs[(100 * i):(100 * (i + 1)), :] = z
40 | 
41 |     indices = np.array([np.where(labels == i)[0] for i in range(10)])
42 |     classes = np.array([zs[index] for index in indices])
43 |     means = np.array([np.mean(c, axis=0) for c in classes])
44 | 
45 |     fig = plt.figure(figsize=(10, 10))
46 |     ax = fig.add_subplot(111)
47 |     ax.scatter(zs[:, 0], zs[:, 1], c=labels)
48 | 
49 |     # annotate means
50 |     for i, mean in enumerate(means):
51 |         ax.annotate(str(i), xy=mean, size=16,
52 |                     bbox=dict(boxstyle='round', facecolor='white', alpha=0.9))
53 |     # plot details
54 |     ax.set_xlim([-5, 5])
55 |     ax.set_ylim([-5, 5])
56 |     ax.set_xticks([])
57 |     ax.set_yticks([])
58 |     plt.tight_layout()
59 | 
60 |     plt.savefig('figs/train/scatter/' + name + '.png', format='png')
61 | 


--------------------------------------------------------------------------------
/providers/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/providers/__init__.py


--------------------------------------------------------------------------------
/providers/anime.py:
--------------------------------------------------------------------------------
 1 | from PIL import Image
 2 | import tensorflow as tf
 3 | import numpy as np
 4 | from tqdm import tqdm
 5 | from utils.path import getfiles
 6 | 
 7 | 
 8 | class Anime(object):
 9 |     def __init__(self, image_folder=None, npy='faces/faces.npy'):
10 |         try:
11 |             self.data = np.load(npy)
12 |         except:
13 |             self.data = self._get_data(image_folder)
14 |             self.save_txt(npy)
15 | 
16 |     def _get_data(self, image_folder):
17 | 
18 |         image_names = getfiles(image_folder)
19 | 
20 |         images = []
21 | 
22 |         for image_name in tqdm(image_names):
23 |             image = np.array(Image.open('{0}/{1}'.format(image_folder, image_name)))
24 |             if np.ndim(image) == 3:
25 |                 image = (image / 255.) * 2 - 1
26 |                 image = np.pad(image, ((2, 2), (2, 2), (0, 0)), mode='constant')
27 |                 images.append(image.flatten())
28 | 
29 |         images = np.array(images)
30 | 
31 |         return images
32 | 
33 |     def save_txt(self, path='faces/faces.npy'):
34 |         np.save(path, self.data)
35 | 
36 |     def get_size(self):
37 |         return self.data.shape[0]
38 | 
39 |     def next_batch(self, num):
40 |         '''
41 |         Return a total of `num` random samples and labels.
42 |         '''
43 |         idx = np.arange(0, self.get_size())
44 |         np.random.shuffle(idx)
45 |         idx = idx[:num]
46 |         batch = self.data[idx]
47 | 
48 |         return batch


--------------------------------------------------------------------------------
/sample.py:
--------------------------------------------------------------------------------
 1 | from coders.vae_coding import conv_anime_decoder, conv_anime_encoder
 2 | import tensorflow as tf
 3 | import numpy as np
 4 | from plots.grid_plots import show_samples, show_latent_scatter
 5 | from providers.anime import Anime
 6 | from tqdm import tqdm
 7 | from models.generator import GENERATOR
 8 | 
 9 | def main():
10 |     flags = tf.flags
11 |     flags.DEFINE_integer("latent_dim", 64, "Dimension of latent space.")
12 |     flags.DEFINE_integer("obs_dim", 12288, "Dimension of observation space.")
13 |     flags.DEFINE_integer("batch_size", 60, "Batch size.")
14 |     flags.DEFINE_integer("epochs", 500, "As it said")
15 |     flags.DEFINE_integer("updates_per_epoch", 100, "Really just can set to 1 if you don't like mini-batch.")
16 |     FLAGS = flags.FLAGS
17 | 
18 |     kwargs = {
19 |         'latent_dim': FLAGS.latent_dim,
20 |         'observation_dim': FLAGS.obs_dim,
21 |         'generator': conv_anime_decoder,
22 |         'obs_distrib': 'Gaussian'
23 |     }
24 |     g = GENERATOR(**kwargs)
25 |     g.load_pretrained("weights/vae_anime/generator")
26 | 
27 |     z = np.random.normal(size=[FLAGS.batch_size, FLAGS.latent_dim])
28 |     samples = g.e2x(z)
29 |     print samples.shape
30 |     show_samples(samples, 4, 15, [64, 64, 3], name='small_samples', shift=True)
31 | 
32 | if __name__ == '__main__':
33 |     main()


--------------------------------------------------------------------------------
/utils/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/utils/__init__.py


--------------------------------------------------------------------------------
/utils/path.py:
--------------------------------------------------------------------------------
 1 | import os
 2 | import pandas as pd
 3 | from numpy import genfromtxt
 4 | 
 5 | 
 6 | def pathfinder(path):
 7 |     script_dir = os.path.dirname('__file__')
 8 |     fullpath = os.path.join(script_dir, path)
 9 |     return fullpath
10 | 
11 | 
12 | def readdata(path, pandas=False):
13 |     fullpath = pathfinder(path)
14 |     if not pandas:
15 |         return genfromtxt(fullpath, delimiter=',')
16 |     else:
17 |         return pd.read_csv(fullpath)
18 | 
19 | 
20 | def getfiles(path):
21 |     files = []
22 |     for (dirpath, dirnames, filenames) in os.walk(path):
23 |         files.extend(filenames)
24 |     return files


--------------------------------------------------------------------------------
/vae_train.py:
--------------------------------------------------------------------------------
 1 | from coders.vae_coding import fc_mnist_encoder, fc_mnist_decoder
 2 | import tensorflow as tf
 3 | import numpy as np
 4 | from plots.grid_plots import show_samples, show_latent_scatter
 5 | from tensorflow.examples.tutorials.mnist import input_data
 6 | from tqdm import tqdm
 7 | from models.vae import VAE
 8 | 
 9 | """
10 | This simple implementation is heavily refer on some github code online.
11 | Such as:
12 | https://github.com/kvfrans/variational-autoencoder
13 | https://github.com/hwalsuklee/tensorflow-mnist-VAE
14 | etc
15 | 
16 | The entire purpose of releasing this code is to help people understand the simple structure of VAE.
17 | """
18 | 
19 | def main():
20 |     flags = tf.flags
21 |     flags.DEFINE_integer("latent_dim", 2, "Dimension of latent space.")
22 |     flags.DEFINE_integer("batch_size", 128, "Batch size.")
23 |     flags.DEFINE_integer("epochs", 500, "As it said")
24 |     flags.DEFINE_integer("updates_per_epoch", 100, "Really just can set to 1 if you don't like mini-batch.")
25 |     flags.DEFINE_string("data_dir", 'mnist', "Tensorflow demo data download position.")
26 |     FLAGS = flags.FLAGS
27 | 
28 |     kwargs = {
29 |         'latent_dim': FLAGS.latent_dim,
30 |         'batch_size': FLAGS.batch_size,
31 |         'encoder': fc_mnist_encoder,
32 |         'decoder': fc_mnist_decoder
33 |     }
34 |     vae = VAE(**kwargs)
35 |     mnist = input_data.read_data_sets(train_dir=FLAGS.data_dir)
36 |     tbar = tqdm(range(FLAGS.epochs))
37 |     for epoch in tbar:
38 |         training_loss = 0.
39 | 
40 |         for _ in range(FLAGS.updates_per_epoch):
41 |             x, _ = mnist.train.next_batch(FLAGS.batch_size)
42 |             loss = vae.update(x)
43 |             training_loss += loss
44 | 
45 |         training_loss /= FLAGS.updates_per_epoch
46 |         s = "Loss: {:.4f}".format(training_loss)
47 |         tbar.set_description(s)
48 | 
49 |     z = np.random.normal(size=[FLAGS.batch_size, FLAGS.latent_dim])
50 |     samples = vae.z2x(z)[0]
51 |     show_samples(samples, 10, 10, [28, 28], name='samples')
52 |     show_latent_scatter(vae, mnist, name='latent')
53 | 
54 |     vae.save_generator('weights/vae_mnist/generator')
55 | 
56 | 
57 | if __name__ == '__main__':
58 |     main()


--------------------------------------------------------------------------------
/vae_train_anime.py:
--------------------------------------------------------------------------------
 1 | from coders.vae_coding import conv_anime_decoder, conv_anime_encoder
 2 | import tensorflow as tf
 3 | import numpy as np
 4 | from plots.grid_plots import show_samples, show_latent_scatter
 5 | from providers.anime import Anime
 6 | from tqdm import tqdm
 7 | from models.vae import VAE
 8 | 
 9 | 
10 | def main():
11 |     flags = tf.flags
12 |     flags.DEFINE_integer("latent_dim", 64, "Dimension of latent space.")
13 |     flags.DEFINE_integer("obs_dim", 12288, "Dimension of observation space.")
14 |     flags.DEFINE_integer("batch_size", 64, "Batch size.")
15 |     flags.DEFINE_integer("epochs", 500, "As it said")
16 |     flags.DEFINE_integer("updates_per_epoch", 100, "Really just can set to 1 if you don't like mini-batch.")
17 |     FLAGS = flags.FLAGS
18 | 
19 |     kwargs = {
20 |         'latent_dim': FLAGS.latent_dim,
21 |         'batch_size': FLAGS.batch_size,
22 |         'observation_dim': FLAGS.obs_dim,
23 |         'encoder': conv_anime_encoder,
24 |         'decoder': conv_anime_decoder,
25 |         'observation_distribution': 'Gaussian'
26 |     }
27 |     vae = VAE(**kwargs)
28 |     provider = Anime()
29 |     tbar = tqdm(range(FLAGS.epochs))
30 |     for epoch in tbar:
31 |         training_loss = 0.
32 | 
33 |         for _ in range(FLAGS.updates_per_epoch):
34 |             x = provider.next_batch(FLAGS.batch_size)
35 |             loss = vae.update(x)
36 |             training_loss += loss
37 | 
38 |         training_loss /= FLAGS.updates_per_epoch
39 |         s = "Loss: {:.4f}".format(training_loss)
40 |         tbar.set_description(s)
41 | 
42 |     z = np.random.normal(size=[FLAGS.batch_size, FLAGS.latent_dim])
43 |     samples = vae.z2x(z)[0]
44 |     show_samples(samples, 8, 8, [64, 64, 3], name='samples')
45 | 
46 |     vae.save_generator('weights/vae_anime/generator')
47 | 
48 | 
49 | if __name__ == '__main__':
50 |     main()


--------------------------------------------------------------------------------
/weights/vae_anime/checkpoint:
--------------------------------------------------------------------------------
1 | model_checkpoint_path: "generator"
2 | all_model_checkpoint_paths: "generator"
3 | 


--------------------------------------------------------------------------------
/weights/vae_anime/generator.data-00000-of-00001:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_anime/generator.data-00000-of-00001


--------------------------------------------------------------------------------
/weights/vae_anime/generator.index:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_anime/generator.index


--------------------------------------------------------------------------------
/weights/vae_anime/generator.meta:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_anime/generator.meta


--------------------------------------------------------------------------------
/weights/vae_mnist/checkpoint:
--------------------------------------------------------------------------------
1 | model_checkpoint_path: "generator"
2 | all_model_checkpoint_paths: "generator"
3 | 


--------------------------------------------------------------------------------
/weights/vae_mnist/generator.data-00000-of-00001:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_mnist/generator.data-00000-of-00001


--------------------------------------------------------------------------------
/weights/vae_mnist/generator.index:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_mnist/generator.index


--------------------------------------------------------------------------------
/weights/vae_mnist/generator.meta:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/wuga214/IMPLEMENTATION_Variational-Auto-Encoder/6fb103bd44e2887b827861d5026f054802a76bd0/weights/vae_mnist/generator.meta


--------------------------------------------------------------------------------