modelit 0.2.2__tar.gz → 0.2.4__tar.gz

modelit-0.2.4/CONTRIBUTING.md

@@ -0,0 +1,53 @@
+# Contributing to ModelIt
+
+Thanks for helping improve ModelIt.
+
+## Add a new template
+
+1. Fork this repo.
+2. Create a new folder:
+
+```text
+modelit/templates/<name>/
+```
+
+3. Add your code in:
+
+```text
+modelit/templates/<name>/template.py
+```
+
+4. Make sure the folder name is the function name.
+5. Test it locally:
+
+```python
+from modelit import <name>
+
+<name>()
+```
+
+or:
+
+```bash
+modelit create <name>
+```
+
+## Rules
+
+- Keep templates simple and runnable.
+- Avoid internet-only dependencies.
+- Do not add `metadata.json`.
+- Keep file names lowercase.
+- Prefer clean, beginner-friendly code.
+
+## Pull request flow
+
+1. Fork the repo.
+2. Create a branch.
+3. Add your template.
+4. Commit and push.
+5. Open a pull request.
+
+## Maintainer notes
+
+I will review templates for clarity, correctness, and portability before merging.

{modelit-0.2.2 → modelit-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modelit
-Version: 0.2.2
+Version: 0.2.4
 Summary: Local-first ML starter templates you can print or save
 Author: Yashsmith
 License-Expression: MIT

modelit-0.2.4/modelit/templates/cnn/template.py

@@ -0,0 +1,491 @@
+# =========================================================
+# CNN FROM SCRATCH + CNN USING KERAS
+# Uses Built-in MNIST Dataset
+# No Kaggle Download Needed
+# =========================================================
+
+
+# =========================================================
+# IMPORTS
+# =========================================================
+
+import cv2
+import numpy as np
+import matplotlib.pyplot as plt
+
+from tensorflow.keras.datasets import mnist
+from tensorflow.keras.models import Sequential
+
+from tensorflow.keras.layers import (
+    Conv2D,
+    MaxPooling2D,
+    Flatten,
+    Dense
+)
+
+from sklearn.model_selection import train_test_split
+
+
+# =========================================================
+# LOAD SAMPLE IMAGE
+# =========================================================
+
+(X_train_full, y_train_full), (X_test_full, y_test_full) = mnist.load_data()
+
+image = X_train_full[0]
+
+print("Original Image Shape:", image.shape)
+
+plt.imshow(image, cmap="gray")
+plt.title("Original Image")
+plt.axis("off")
+plt.show()
+
+
+# =========================================================
+# CONVOLUTION FROM SCRATCH
+# =========================================================
+
+def convolution(image, kernel, strides=1):
+
+    image_height, image_width = image.shape
+
+    kernel_height, kernel_width = kernel.shape
+
+    output_height = (
+        (image_height - kernel_height) // strides
+    ) + 1
+
+    output_width = (
+        (image_width - kernel_width) // strides
+    ) + 1
+
+    output = np.zeros((output_height, output_width))
+
+    for i in range(1, image.shape[0] - 1, strides):
+
+        for j in range(1, image.shape[1] - 1, strides):
+
+            val = (
+
+                image[i - 1, j - 1] * kernel[0, 0] +
+
+                image[i - 1, j] * kernel[0, 1] +
+
+                image[i - 1, j + 1] * kernel[0, 2] +
+
+                image[i, j - 1] * kernel[1, 0] +
+
+                image[i, j] * kernel[1, 1] +
+
+                image[i, j + 1] * kernel[1, 2] +
+
+                image[i + 1, j - 1] * kernel[2, 0] +
+
+                image[i + 1, j] * kernel[2, 1] +
+
+                image[i + 1, j + 1] * kernel[2, 2]
+
+            )
+
+            out_i = (i - 1) // strides
+
+            out_j = (j - 1) // strides
+
+            output[out_i, out_j] = val
+
+    return output
+
+
+# =========================================================
+# EDGE DETECTION KERNEL
+# =========================================================
+
+kernel = np.array([
+    [-1, -1, -1],
+    [-1,  8, -1],
+    [-1, -1, -1]
+])
+
+
+# =========================================================
+# APPLY CONVOLUTION
+# =========================================================
+
+conv_img = convolution(image, kernel, 1)
+
+plt.imshow(conv_img, cmap="gray")
+plt.title("Convolution Output")
+plt.axis("off")
+plt.show()
+
+
+# =========================================================
+# RELU ACTIVATION
+# =========================================================
+
+def relu(image):
+
+    for i in range(0, image.shape[0]):
+
+        for j in range(0, image.shape[1]):
+
+            if image[i, j] >= 0:
+
+                image[i, j] = image[i, j]
+
+            else:
+
+                image[i, j] = 0
+
+    return image
+
+
+# =========================================================
+# APPLY RELU
+# =========================================================
+
+relu_img = relu(conv_img)
+
+plt.imshow(relu_img, cmap="gray")
+plt.title("ReLU Output")
+plt.axis("off")
+plt.show()
+
+
+# =========================================================
+# MAX POOLING
+# =========================================================
+
+def max_pooling(image, pool_size=2, strides=2):
+
+    image_height, image_width = image.shape
+
+    output_height = (
+        (image_height - pool_size) // strides
+    ) + 1
+
+    output_width = (
+        (image_width - pool_size) // strides
+    ) + 1
+
+    output = np.zeros((output_height, output_width))
+
+    for i in range(0, image_height - 1, strides):
+
+        for j in range(0, image_width - 1, strides):
+
+            val = image[i, j]
+
+            if image[i, j + 1] > val:
+
+                val = image[i, j + 1]
+
+            if image[i + 1, j] > val:
+
+                val = image[i + 1, j]
+
+            if image[i + 1, j + 1] > val:
+
+                val = image[i + 1, j + 1]
+
+            out_i = i // strides
+
+            out_j = j // strides
+
+            if (
+                out_i < output_height and
+                out_j < output_width
+            ):
+
+                output[out_i, out_j] = val
+
+    return output
+
+
+# =========================================================
+# APPLY MAX POOLING
+# =========================================================
+
+pool_img = max_pooling(relu_img, 2, 2)
+
+plt.imshow(pool_img, cmap="gray")
+plt.title("Max Pooling Output")
+plt.axis("off")
+plt.show()
+
+
+# =========================================================
+# FLATTEN
+# =========================================================
+
+flat_img = pool_img.flatten()
+
+print("\n=================================================")
+
+print("Original Image Shape:", image.shape)
+
+print("Convolution Shape:", conv_img.shape)
+
+print("ReLU Shape:", relu_img.shape)
+
+print("Pooling Shape:", pool_img.shape)
+
+print("Flatten Shape:", flat_img.shape)
+
+print("=================================================")
+
+
+# =========================================================
+# DENSE LAYER
+# =========================================================
+
+weights = np.random.rand(flat_img.shape[0])
+
+dense_output = np.dot(flat_img, weights)
+
+print("\nDense Layer Output:", dense_output)
+
+
+# =========================================================
+# CNN USING KERAS
+# =========================================================
+
+print("\n=================================================")
+print("CNN USING KERAS")
+print("=================================================")
+
+
+# =========================================================
+# NORMALIZATION
+# =========================================================
+
+X_train_full = X_train_full / 255.0
+
+X_test_full = X_test_full / 255.0
+
+
+# =========================================================
+# RESHAPE FOR CNN
+# =========================================================
+
+X_train_full = X_train_full.reshape(-1, 28, 28, 1)
+
+X_test_full = X_test_full.reshape(-1, 28, 28, 1)
+
+
+# =========================================================
+# TRAIN TEST SPLIT
+# =========================================================
+
+X_train, X_val, y_train, y_val = train_test_split(
+
+    X_train_full,
+    y_train_full,
+
+    test_size=0.2,
+
+    random_state=42
+
+)
+
+
+print("\nTraining Samples:", X_train.shape[0])
+
+print("Validation Samples:", X_val.shape[0])
+
+
+# =========================================================
+# CREATE CNN MODEL
+# =========================================================
+
+model = Sequential()
+
+
+# =========================================================
+# CONVOLUTION + RELU + POOLING
+# =========================================================
+
+model.add(
+
+    Conv2D(
+        32,
+        (3, 3),
+        activation="relu",
+        input_shape=(28, 28, 1)
+    )
+
+)
+
+model.add(
+
+    MaxPooling2D((2, 2))
+
+)
+
+
+# =========================================================
+# SECOND CNN BLOCK
+# =========================================================
+
+model.add(
+
+    Conv2D(
+        64,
+        (3, 3),
+        activation="relu"
+    )
+
+)
+
+model.add(
+
+    MaxPooling2D((2, 2))
+
+)
+
+
+# =========================================================
+# THIRD CNN BLOCK
+# =========================================================
+
+model.add(
+
+    Conv2D(
+        128,
+        (3, 3),
+        activation="relu"
+    )
+
+)
+
+model.add(
+
+    MaxPooling2D((2, 2))
+
+)
+
+
+# =========================================================
+# FLATTEN
+# =========================================================
+
+model.add(Flatten())
+
+
+# =========================================================
+# DENSE LAYER
+# =========================================================
+
+model.add(
+
+    Dense(
+        64,
+        activation="relu"
+    )
+
+)
+
+
+# =========================================================
+# OUTPUT LAYER
+# MNIST = 10 CLASSES
+# =========================================================
+
+model.add(
+
+    Dense(
+        10,
+        activation="softmax"
+    )
+
+)
+
+
+# =========================================================
+# MODEL SUMMARY
+# =========================================================
+
+model.summary()
+
+
+# =========================================================
+# COMPILE MODEL
+# =========================================================
+
+model.compile(
+
+    optimizer="adam",
+
+    loss="sparse_categorical_crossentropy",
+
+    metrics=["accuracy"]
+
+)
+
+
+# =========================================================
+# TRAIN MODEL
+# =========================================================
+
+train_model = model.fit(
+
+    X_train,
+    y_train,
+
+    epochs=5,
+
+    validation_data=(X_val, y_val)
+
+)
+
+
+# =========================================================
+# EVALUATE MODEL
+# =========================================================
+
+loss, accuracy = model.evaluate(
+    X_test_full,
+    y_test_full
+)
+
+print("\n=================================================")
+
+print("Test Accuracy:", accuracy)
+
+print("Test Loss:", loss)
+
+print("=================================================")
+
+
+# =========================================================
+# PREDICTION
+# =========================================================
+
+prediction = model.predict(
+    X_test_full[0].reshape(1, 28, 28, 1)
+)
+
+predicted_class = np.argmax(prediction)
+
+print("\nPredicted Digit:", predicted_class)
+
+print("Actual Digit:", y_test_full[0])
+
+
+# =========================================================
+# SHOW TEST IMAGE
+# =========================================================
+
+plt.imshow(
+    X_test_full[0].reshape(28, 28),
+    cmap="gray"
+)
+
+plt.title(
+    f"Predicted: {predicted_class}"
+)
+
+plt.axis("off")
+
+plt.show()

{modelit-0.2.2 → modelit-0.2.4}/modelit.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modelit
-Version: 0.2.2
+Version: 0.2.4
 Summary: Local-first ML starter templates you can print or save
 Author: Yashsmith
 License-Expression: MIT

{modelit-0.2.2 → modelit-0.2.4}/modelit.egg-info/SOURCES.txt

@@ -1,4 +1,5 @@
 .gitignore
+CONTRIBUTING.md
 LICENSE
 MANIFEST.in
 README.md
@@ -14,5 +15,6 @@ modelit.egg-info/SOURCES.txt
 modelit.egg-info/dependency_links.txt
 modelit.egg-info/entry_points.txt
 modelit.egg-info/top_level.txt
-modelit/templates/backpropogation/template.py
+modelit/templates/backpropagation/template.py
+modelit/templates/cnn/template.py
 modelit/templates/perceptron/template.py