vault backup: 2023-06-07 15:02:38

.obsidian/workspace.json

@@ -36,13 +36,13 @@
             }
           },
           {
-            "id": "a4bd0e4c3fda45c7",
+            "id": "3971d194ef2e5886",
             "type": "leaf",
             "state": {
               "type": "markdown",
               "state": {
-                "file": "04. Programming/Python/argparse.ArgumentParser.md",
-                "mode": "source",
+                "file": "04. Programming/Machine Learning.md",
+                "mode": "preview",
                 "source": true
               }
             }
@@ -114,7 +114,7 @@
             "state": {
               "type": "backlink",
               "state": {
-                "file": "04. Programming/Python/argparse.ArgumentParser.md",
+                "file": "04. Programming/Machine Learning.md",
                 "collapseAll": false,
                 "extraContext": false,
                 "sortOrder": "alphabetical",
@@ -139,7 +139,7 @@
             "state": {
               "type": "outline",
               "state": {
-                "file": "04. Programming/Python/argparse.ArgumentParser.md"
+                "file": "04. Programming/Machine Learning.md"
               }
             }
           },
@@ -190,8 +190,12 @@
       "periodic-notes:Open today": false
     }
   },
-  "active": "a4bd0e4c3fda45c7",
+  "active": "3971d194ef2e5886",
   "lastOpenFiles": [
+    "04. Programming/categorical_crossentropy.md",
+    "04. Programming/Machine Learning.md",
+    "00. Inbox/01. TODO.md",
+    "04. Programming/Python/argparse.ArgumentParser.md",
     "05. 資料收集/皮質醇.md",
     "05. 資料收集/讀書筆記/20230206 - 卡片盒筆記.md",
     "05. 資料收集/讀書筆記/20201201 - 學習如何學習.md",
@@ -211,18 +215,14 @@
     "05. 資料收集/讀書筆記/20220619 - 精確的力量.md",
     "05. 資料收集/讀書筆記/20220526 - 深入淺出設計模式.md",
     "04. Programming/Python/matplotlib.md",
-    "04. Programming/Machine Learning.md",
     "05. 資料收集/核心肌群訓練.md",
     "05. 資料收集/Capture One.md",
     "05. 資料收集/AI.md",
     "00. Inbox/My Mindmap.canvas",
     "00. Inbox/Habit Tracker.md",
     "00. Inbox/00. Inbox.md",
-    "04. Programming/Python/argparse.ArgumentParser.md",
-    "00. Inbox/01. TODO.md",
     "01. 個人/01. Daily/2023-05-11(週四).md",
     "01. 個人/01. Daily/2023-05-12(週五).md",
-    "01. 個人/01. Daily/2023-05-10(週三).md",
     "attachments/Pasted image 20230504183452.png",
     "attachments/Pasted image 20230504183439.png",
     "00. Inbox/想要的鏡頭/變焦",

04. Programming/Machine Learning.md

@@ -4,7 +4,7 @@
 |  問題類型   | 啟動函數 |                損失函數                |
 |:-----------:|:--------:|:--------------------------------------:|
 |  二元分類   | sigmoid  |   binary_crossentropy（二元交叉熵）    | 
-|  多元分類   | softmax  | categorical_crossentropy（分類交叉熵） |
+|  單標籤多元分類   | softmax  | [[categorical_crossentropy]]（分類交叉熵）<br> sparse_categorical_crossentropy |
 | 多標籤分類  | sigmoid  |          binary_crossentropy           |
 |  回歸求值   |   None   |            mse（均方誤差）             |
 | 回歸求0~1值 | sigmoid  |        mse或binary_crossentropy        |

04. Programming/categorical_crossentropy.md

@@ -0,0 +1,2 @@
+- 僅適用於 one-hot 編碼。
+- 如果輸出不是 one-hot，而是整數標籤，也就是直接輸出 0、1、2，而不是一個array（`[0, 0, 0, 1, 0]` 之類），那就需要 sparse_categorical_crossentropy。