Method fit

code/chapter6.py:345–387

参数说明： X_train：训练数据 y_train：训练数据标签 n_epochs：epoch 次数 batch_size：每次 epoch 的 batch size verbose：是否每个 batch 输出损失 epo_verbose：是否每个 epoch 输出损失

(self, X_train, y_train, n_epochs=20, batch_size=64, verbose=False, epo_verbose=True)

            v.flush_gradients()
    
    def fit(self, X_train, y_train, n_epochs=20, batch_size=64, verbose=False, epo_verbose=True):
        """
        参数说明：
        X_train：训练数据
        y_train：训练数据标签
        n_epochs：epoch 次数
        batch_size：每次 epoch 的 batch size
        verbose：是否每个 batch 输出损失
        epo_verbose：是否每个 epoch 输出损失
        """
        self.verbose = verbose
        self.n_epochs = n_epochs
        self.batch_size = batch_size
        
        if not self.is_initialized:
            self.n_features = X_train.shape[1]
            self._set_params()
        
        prev_loss = np.inf
        for i in range(n_epochs):
            loss, epoch_start = 0.0, time.time()
            batch_generator, n_batch = minibatch(X_train, self.batch_size, shuffle=True)

            for j, batch_idx in enumerate(batch_generator):
                batch_len, batch_start = len(batch_idx), time.time()
                X_batch, y_batch = X_train[batch_idx], y_train[batch_idx]
                out, _ = self.forward(X_batch)
                y_pred_batch = softmax(out)
                batch_loss = self.loss(y_batch, y_pred_batch)
                grad = self.loss.grad(y_batch, y_pred_batch)
                _, _ = self.backward(grad)
                self.update()
                loss += batch_loss

                if self.verbose:
                    fstr = "\t[Batch {}/{}] Train loss: {:.3f} ({:.1f}s/batch)"
                    print(fstr.format(j + 1, n_batch, batch_loss, time.time() - batch_start))

            loss /= n_batch
            if epo_verbose:
                fstr = "[Epoch {}] Avg. loss: {:.3f}  Delta: {:.3f} ({:.2f}m/epoch)"
                print(fstr.format(i + 1, loss, prev_loss - loss, (time.time() - epoch_start) / 60.0))
            prev_loss = loss
            
    def evaluate(self, X_test, y_test, batch_size=128):
        acc = 0.0