Merge pull request #18 from shenxiangzhuang/doc/improve

Docs:improve

.github/workflows/test.yml

@@ -13,7 +13,7 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest]
-        python-version: ["3.8", "3.9"]
+        python-version: ["3.8", "3.9", "3.10", "3.11"]
 
     steps:
       - name: Checkout
@@ -51,7 +51,7 @@ jobs:
 
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v3
-        if: success() && matrix.os == 'ubuntu-latest' && matrix.python-version == '3.9'
+        if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.8' && success()
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           files: ./coverage.xml

docs/index.md

@@ -2,6 +2,12 @@
 There are some deep learning algorithms implemented from scratch!
 Let's learn deep learning with toy code.
 
+The implementation of toydl is highly inspired by [minitorch](https://github.com/minitorch/minitorch).
+
+
 ## Overview
 
-- [ ] Mini-Torch
+ToyDl, as an education project, is aimed to make the
+concepts behind the deep learning as clear as possible.
+I do believe the key to reach the target is **SIMPLE**, as simple as possible.
+I'm trying to use little code to build this library, although it's not simple enough yet.

docs/quickstart.md

@@ -1 +1,67 @@
 # Quick Start
+
+
+
+## Binary Classification with MLP
+
+??? MLPBinaryClassifyModel
+
+    ```python
+    from toydl.core.optim import SGD
+    from toydl.core.scalar import Scalar
+    from toydl.network.mlp import MLPBinaryClassify, MLPConfig
+
+
+    class MLPBinaryClassifyModel:
+        def __init__(self, mlp_config: MLPConfig):
+            self.model = MLPBinaryClassify(mlp_config)
+
+        def train(self, data, learning_rate, max_epochs=500):
+            optim = SGD(self.model.parameters(), learning_rate)
+
+            losses = []
+            for epoch in range(1, max_epochs + 1):
+                total_loss = 0.0
+                correct = 0
+                optim.zero_grad()
+
+                # Forward
+                for i in range(data.n):
+                    x_1, x_2 = data.X[i]
+                    y = data.y[i]
+                    x_1 = Scalar(x_1)
+                    x_2 = Scalar(x_2)
+                    out = self.model.forward((x_1, x_2))
+                    if y == 1:
+                        prob = out
+                        correct += 1 if out.data > 0.5 else 0
+                    else:
+                        prob = -out + 1.0
+                        correct += 1 if out.data < 0.5 else 0
+                    loss = -prob.log()
+                    (loss / data.n).backward()
+                    total_loss += loss.data
+
+                losses.append(total_loss)
+
+                # Update
+                optim.step()
+
+                # Logging
+                if epoch % 10 == 0 or epoch == max_epochs:
+                    print("Epoch ", epoch, " loss ", total_loss, "correct", correct)
+
+
+    if __name__ == "__main__":
+        from toydl.dataset.simulation import datasets
+
+        PTS = 50
+        RATE = 0.05
+        data = datasets["Simple"](PTS)
+
+        config = MLPConfig(in_size=2, out_size=1, hidden_layer_size=5, hidden_layer_num=2)
+        mlp_model = MLPBinaryClassifyModel(config)
+        # print(mlp_model.model.order_layer_names)
+        mlp_model.train(data, RATE, max_epochs=500)
+
+    ```

mkdocs.yml

@@ -11,8 +11,6 @@ nav:
   - Home: index.md
   - Getting started:
       - Quick Start: quickstart.md
-      - User Guide:
-          - Read First: user-guide/1-read-first.md
   - API References:
       - Operator: api/operator.md
       - Context: api/context.md
@@ -29,13 +27,17 @@ theme:
   palette:
     - media: "(prefers-color-scheme: light)"
       scheme: default
+      primary: teal
+      accent: deep purple
       toggle:
-        icon: material/lightbulb-outline
+        icon: material/weather-sunny
         name: Switch to dark mode
     - media: "(prefers-color-scheme: dark)"
       scheme: slate
+      primary: cyan
+      accent: deep purple
       toggle:
-        icon: material/lightbulb
+        icon: material/weather-night
         name: Switch to light mode
   features:
     # - announce.dismiss