更新逻辑

Author: ctkqiang

.vscode/launch.json

@@ -0,0 +1,16 @@
+{
+    // 使用 IntelliSense 了解相关属性。 
+    // 悬停以查看现有属性的描述。
+    // 欲了解更多信息，请访问: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python: 当前文件",
+            "type": "python",
+            "request": "launch",
+            "program": "${file}",
+            "console": "integratedTerminal",
+            "justMyCode": true
+        }
+    ]
+}

Readme.md

@@ -115,7 +115,11 @@ HuaTuoAI/
 ### 用法
 
 ```bash
-python3 main.py
+python3 tensorflow/run.py
+```
+或
+```bash
+python3 pytorch/run.py
 ```
 
 ### 以下是使用 chinese_medicine.h5 二进制文件进行程序演示的示例，以及输出结果:

main.py old_run.pymain.py renamed to old_run.py

@@ -19,12 +19,15 @@
     from keras.layers import Conv2D, MaxPooling2D
     from keras.layers import Activation, Dropout, Flatten, Dense
     from keras import backend as _keras
-except:
+except ImportError:
     raise "🥹无法安装配件"
-
+finally:
+    pass
 
 class HuaTuoAI:
     def __init__(self):
+        assert tf.__version__.startswith("2")
+
         self.chinese_medicine_url: str = "https://raw.githubusercontent.com/johnmelodyme/HuaTuoAI/main/data/chinese_medicine.txt"
         self.image_data: str = "https://github.com/johnmelodyme/HuaTuoAI/releases/download/images/images.zip"
         self.image_width: int = 224
@@ -132,6 +135,8 @@ def train(self) -> None:
         )
 
         try:
+       
+
             self.model.save_weights("chinese_medicine{}".format(self.binary_extension))
             self.log(msg="😇训练模型保存为《chinese_medicine.h5》")
             self.convert_to_tflite()

run.py

@@ -0,0 +1,106 @@
+# -*- coding: UTF-8 -*-
+try:
+    import os
+    import tensorflow as tf
+    import numpy as np
+    import matplotlib
+    import matplotlib.pyplot as 图表
+    from typing import Tuple
+    from datetime import datetime
+except ImportError:
+    raise "🥹无法安装配件"
+finally:
+    pass
+
+class 华佗AI:
+    def __init__(self) -> None:
+        super(华佗AI, self).__init__()
+        
+        matplotlib.rcParams["font.family"] = "Heiti TC"
+        
+        self.图片文件 = "./data/images/"
+        self.seed :int = 21
+        self.图像高度 :int = 32
+        self.图像宽度 :int = 32
+        self.批量的大小 :int = 20 
+        self.轴 :str | None = "off" 
+        self.活化 :str | None = "relu"
+        self.中药材料 :list | any = [
+            "丁公藤", 
+            "金银花",
+            "罗汉果",
+            "人参片"
+        ]
+        self.优化器 :str | None = "adam"
+        self.纪元 :int = 10
+        
+    def 训练数据(self) -> None:
+        self.记录(信息="TensorFlow 版本 |> {}".format(tf.__version__))
+        
+        if os.path.exists(self.图片文件):
+            训练目录 :tensorflow.python.data.ops.batch_op._BatchDataset = tf.keras.utils.image_dataset_from_directory(
+                self.图片文件,
+                image_size=(self.图像高度, self.图像宽度),
+                batch_size=self.批量的大小
+            )
+            
+            估价目录 :tensorflow.python.data.ops.batch_op._BatchDataset = tf.keras.utils.image_dataset_from_directory(
+                self.图片文件,
+                image_size=(self.图像高度, self.图像宽度),
+                batch_size=self.批量的大小
+            )
+            
+            测试目录 :tensorflow.python.data.ops.batch_op._BatchDataset = tf.keras.utils.image_dataset_from_directory(
+                self.图片文件,
+                image_size=(self.图像高度, self.图像宽度),
+                batch_size=self.批量的大小
+            )
+            
+            图表.figure(figsize=(10, 10))
+            
+
+            for 照片, 名称 in 训练目录.take(1):
+                for 药 in range(9):
+                    图表.subplot(3, 3, (药 + 1))
+                    图表.imshow(照片[药].numpy().astype("uint8"))
+                    图表.title(label=self.中药材料[名称[药]], fontsize=10, color="green")
+                    图表.axis(self.轴)
+            
+            训练模型 = tf.keras.Sequential(
+                [
+                    tf.keras.layers.Rescaling(1./255),
+                    tf.keras.layers.Conv2D(32, 3, activation=self.活化),
+                    tf.keras.layers.MaxPool2D(),
+                    tf.keras.layers.Conv2D(32, 3, activation=self.活化),
+                    tf.keras.layers.MaxPool2D(),
+                    tf.keras.layers.Conv2D(32, 3, activation=self.活化),
+                    tf.keras.layers.MaxPool2D(),
+                    tf.keras.layers.Flatten(),
+                    tf.keras.layers.Dense(128, activation=self.活化),
+                    tf.keras.layers.Dense(2)
+                ]
+            )
+            
+            训练模型.compile(
+                optimizer=self.优化器, 
+                loss=tf.losses.SparseCategoricalCrossentropy(from_logits=True),
+                metrics=["accuracy"]
+            )
+            
+            训练模型.fit(训练目录, validation_data=训练目录, epochs=self.纪元)
+        else:
+            self.记录(信息=f"{self.图片文件} 不存在")
+            return
+    
+    def 加载数据(self) -> None:
+        中药材料 = None
+        
+    def 记录(self, 信息: str) -> None:
+        现在 :object = datetime.now()
+        目前时间 :object = 现在.strftime("%H:%M:%S")
+        
+        print(f"华佗AI[{目前时间}] {str(信息)}")
+        
+if __name__ == "__main__":
+    huatuoAi = 华佗AI()
+    huatuoAi.训练数据()