train.py的修改

src/image_recognition/train.py

@@ -558,45 +558,48 @@ def main(opt, callbacks=Callbacks()):
         if opt.bucket:
             os.system(f'gsutil cp gs://{opt.bucket}/evolve.csv {save_dir}')  # 下载evolve.csv 
             
-        for _ in range(opt.evolve):  # generations to evolve
+
-            if evolve_csv.exists():  # if evolve.csv exists: select best hyps and mutate
+        #进行opt.evolve指定次数的超参数演化
-                # Select parent(s)
+        for _ in range(opt.evolve): 
-                parent = 'single'  # parent selection method: 'single' or 'weighted'
+            if evolve_csv.exists():  # 如果存在演化过程中生成的csv文件，选择最佳超参数并进行变异
+                # 选择父方法
+                parent = 'single'  # 'single' 表示随机选择一个最佳超参数集作为父代；'weighted' 表示加权选择
                 x = np.loadtxt(evolve_csv, ndmin=2, delimiter=',', skiprows=1)
-                n = min(5, len(x))  # number of previous results to consider
+                n = min(5, len(x))  # 加载超参数数据
-                x = x[np.argsort(-fitness(x))][:n]  # top n mutations
+                x = x[np.argsort(-fitness(x))][:n]  # 考虑前n个历史结果
-                w = fitness(x) - fitness(x).min() + 1E-6  # weights (sum > 0)
+                w = fitness(x) - fitness(x).min() + 1E-6  
                 if parent == 'single' or len(x) == 1:
-                    # x = x[random.randint(0, n - 1)]  # random selection
+                    # 随机选择一个超参数集
-                    x = x[random.choices(range(n), weights=w)[0]]  # weighted selection
+                    x = x[random.choices(range(n), weights=w)[0]]  # 权重选择
                 elif parent == 'weighted':
-                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # weighted combination
+                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # 权重组合
+
+                # 变异操作
+                mp, s = 0.8, 0.2  # 变异概率，标准差
 
-                # Mutate
-                mp, s = 0.8, 0.2  # mutation probability, sigma
                 npr = np.random
                 npr.seed(int(time.time()))
-                g = np.array([meta[k][0] for k in hyp.keys()])  # gains 0-1
+                g = np.array([meta[k][0] for k in hyp.keys()])  #
                 ng = len(meta)
                 v = np.ones(ng)
-                while all(v == 1):  # mutate until a change occurs (prevent duplicates)
+                while all(v == 1):  # 进行变异，直到发生改变，防止生成重复的超参数集
                     v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
-                for i, k in enumerate(hyp.keys()):  # plt.hist(v.ravel(), 300)
+                for i, k in enumerate(hyp.keys()):  
-                    hyp[k] = float(x[i + 7] * v[i])  # mutate
+                    hyp[k] = float(x[i + 7] * v[i])  
 
-            # Constrain to limits
+            # 约束超参数在设定的范围内
             for k, v in meta.items():
-                hyp[k] = max(hyp[k], v[1])  # lower limit
+                hyp[k] = max(hyp[k], v[1])  # 下限
-                hyp[k] = min(hyp[k], v[2])  # upper limit
+                hyp[k] = min(hyp[k], v[2])  # 上限
-                hyp[k] = round(hyp[k], 5)  # significant digits
+                hyp[k] = round(hyp[k], 5)  # 5位有效数字
 
-            # Train mutation
+            # 变异训练
             results = train(hyp.copy(), opt, device, callbacks)
 
-            # Write mutation results
+            #记录结果
             print_mutation(results, hyp.copy(), save_dir, opt.bucket)
 
-        # Plot results
+        # 绘制结果
         plot_evolve(evolve_csv)
         LOGGER.info(f'Hyperparameter evolution finished\n'
                     f"Results saved to {colorstr('bold', save_dir)}\n"