add fourth chapter

cn-Book/4.从零开始实现一个用于文本生成的 GPT 模型.md

@@ -409,8 +409,8 @@ class GELU(nn.Module):
 
     def forward(self, x):
         return 0.5 * x * (1 + torch.tanh(
-          torch.sqrt(torch.tensor(2.0 / torch.pi)) *
-          (x + 0.044715 * torch.pow(x, 3))
+            torch.sqrt(torch.tensor(2.0 / torch.pi)) *
+            (x + 0.044715 * torch.pow(x, 3))
         ))
 ```
 
@@ -456,7 +456,7 @@ class FeedForward(nn.Module):
         )
 
 def forward(self, x):
-		return self.layers(x)
+    return self.layers(x)
 ```
 
 如代码所示，FeedForward 模块是一个小型神经网络，由两个线性层和一个 GELU 激活函数组成。在 1.24 亿参数的 GPT 模型中，该模块可以接收批量输入，每个输入 token 是一个 768 维的向量表示。这一嵌入维度大小通过 `GPT_CONFIG_124M` 配置字典中的 `GPT_CONFIG_124M["emb_dim"]` 参数指定。
@@ -539,9 +539,9 @@ class ExampleDeepNeuralNetwork(nn.Module):
             layer_output = layer(x)
             # Check if shortcut can be applied
             if self.use_shortcut and x.shape == layer_output.shape:
-            		x = x + layer_output
+                x = x + layer_output
             else:
-           		 x = layer_output
+                x = layer_output
         return x
 ```
 
@@ -554,7 +554,7 @@ layer_sizes = [3, 3, 3, 3, 3, 1]
 sample_input = torch.tensor([[1., 0., -1.]])
 torch.manual_seed(123) # specify random seed for the initial weights for reproducibility
 model_without_shortcut = ExampleDeepNeuralNetwork(
-		layer_sizes, use_shortcut=False
+    layer_sizes, use_shortcut=False
 )
 ```
 
@@ -607,7 +607,7 @@ layers.4.0.weight has gradient mean of 0.005049646366387606
 ```python
 torch.manual_seed(123)
 model_with_shortcut = ExampleDeepNeuralNetwork(
-		layer_sizes, use_shortcut=True
+    layer_sizes, use_shortcut=True
 )
 print_gradients(model_with_shortcut, sample_input)
 ```
@@ -804,11 +804,11 @@ class GPTModel(nn.Module):
         self.drop_emb = nn.Dropout(cfg["drop_rate"])
 
         self.trf_blocks = nn.Sequential(
-        		*[TransformerBlock(cfg) for _ in range(cfg["n_layers"])])
+            *[TransformerBlock(cfg) for _ in range(cfg["n_layers"])])
 
         self.final_norm = LayerNorm(cfg["emb_dim"])
         self.out_head = nn.Linear(
-        		cfg["emb_dim"], cfg["vocab_size"], bias=False
+            cfg["emb_dim"], cfg["vocab_size"], bias=False
         )
 
     def forward(self, in_idx):
@@ -851,7 +851,7 @@ print(out)
 ```python
 Input batch:
   tensor([[ 6109, 3626, 6100, 345], # token IDs of text 1
- 				  [ 6109, 1110, 6622, 257]]) # token IDs of text 2
+          [ 6109, 1110, 6622, 257]]) # token IDs of text 2
 
 Output shape: torch.Size([2, 4, 50257])
   tensor([[[ 0.3613, 0.4222, -0.0711, ..., 0.3483, 0.4661, -0.2838],