GPT_model

GPT架构总览如下图所示，我们这里简单创建了一个GPT模型，它是ChatGPT的基础架构。

import tiktoken
import torch
import torch.nn as nn
from torch.utils.data import Dataset, DataLoader 

MultiHeadAttention

class MultiHeadAttention(nn.Module):
    def __init__(self, d_in, d_out,
        context_length, dropout, num_heads, qkv_bias=False):
        super().__init__()
        assert (d_out % num_heads == 0), \
            "d_out must divisible by num_heads"
        self.d_out = d_out 
        self.num_heads = num_heads
        self.head_dim = d_out // num_heads
        self.W_query = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.W_key = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.W_value = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.out_proj = nn.Linear(d_out, d_out)
        self.dropout = nn.Dropout(dropout)
        self.register_buffer(
            "mask",
            torch.triu(torch.ones(context_length, context_length),
                diagonal=1))

    def forward(self, x):
        b, num_tokens, d_in = x.shape
        queries = self.W_query(x)
        keys = self.W_key(x)
        values = self.W_value(x)
        keys = keys.view(b, num_tokens, self.num_heads, self.head_dim)
        values = values.view(b, num_tokens, self.num_heads, self.head_dim)
        queries = queries.view(b, num_tokens, self.num_heads, self.head_dim)
        # (b, num_tokens, num_heads, head_dim) -> (b, num_heads, num_tokens, head_dim)
        keys = keys.transpose(1, 2)
        queries = queries.transpose(1, 2)
        values = values.transpose(1, 2)
        # attn_scores: (b, num_heads, num_tokens, num_tokens)
        attn_scores = queries @ keys.transpose(2, 3)
        mask_bool = self.mask.bool()[:num_tokens, :num_tokens]
        attn_scores.masked_fill_(mask_bool, -torch.inf)

        attn_weights = torch.softmax(
            attn_scores / keys.shape[-1]**0.5, dim=-1)
        attn_weights = self.dropout(attn_weights)
        # (b, num_heads, num_tokens, num_tokens) * (b, num_heads, num_tokens, head_dim)
        # (b, num_heads, num_tokens, head_dim) -> (b, num_tokens, num_heads, head_dim)
        context_vec = (attn_weights @ values).transpose(1,2)
        context_vec = context_vec.contiguous().view(
            b, num_tokens, self.d_out
        )
        context_vec = self.out_proj(context_vec)
        return context_vec

GELU激活函数

GELU激活函数在大模型中运用较多，相比RELU，GELU更平滑，有利于减少梯度爆炸或梯度消失的情况。

class GELU(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, x):
        return 0.5 * x * (1 + torch.tanh(
            torch.sqrt(torch.tensor(2.0 / torch.pi)) * 
            (x + 0.44715 * torch.pow(x, 3))))

FeedForward

class FeedForward(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.layers = nn.Sequential(
            nn.Linear(cfg["emb_dim"], 4 * cfg["emb_dim"]),
            GELU(),
            nn.Linear(4 * cfg["emb_dim"], cfg["emb_dim"]))
        
    def forward(self, x):
        return self.layers(x)

LayerNorm

class LayerNorm(nn.Module):
    def __init__(self, emb_dim):
        super().__init__()
        self.eps = 1e-5
        self.scale = nn.Parameter(torch.ones(emb_dim))
        self.shift = nn.Parameter(torch.zeros(emb_dim))
        
    def forward(self, x):
        mean = x.mean(dim=-1, keepdim=True)
        var = x.var(dim=-1, keepdim=True, unbiased=False)
        norm_x = (x - mean) / torch.sqrt(var + self.eps)
        return self.scale * norm_x + self.shift

TransformerBlock

class TransformerBlock(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.att = MultiHeadAttention(
            d_in=cfg["emb_dim"],
            d_out=cfg["emb_dim"],
            context_length=cfg["context_length"],
            num_heads=cfg["n_heads"],
            dropout=cfg["drop_rate"],
            qkv_bias=cfg["qkv_bias"])
        self.ff = FeedForward(cfg)
        self.norm1 = LayerNorm(cfg["emb_dim"])
        self.norm2 = LayerNorm(cfg["emb_dim"])
        self.drop_shortcut = nn.Dropout(cfg["drop_rate"])
        
    def forward(self, x):
        shortcut = x 
        x = self.norm1(x)
        x = self.att(x)
        x = self.drop_shortcut(x)
        x = x + shortcut
        
        shortcut = x 
        x = self.norm2(x)
        x = self.ff(x)
        x = self.drop_shortcut(x)
        x = x + shortcut
        return x 

GPT

class GPTModel(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.tok_emb = nn.Embedding(cfg["vocab_size"], cfg["emb_dim"])
        self.pos_emb = nn.Embedding(cfg["context_length"], cfg["emb_dim"])
        self.drop_emb = nn.Dropout(cfg["drop_rate"])
        
        self.trf_blocks = nn.Sequential(
            *[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
        )
        
    def forward(self, in_idx):
        batch_size, seq_len = in_idx.shape
        tok_embeds = self.tok_emb(in_idx)
        pos_embeds = self.pos_emb(
            torch.arange(seq_len, device=in_idx.device)
        )
        x = tok_embeds + pos_embeds
        x = self.drop_emb(x)
        x = self.trf_blocks(x)
        x = self.final_norm(x)
        logits = self.out_head(x)
        return logits

GPT_CONFIG_124M = {
    "vocab_size": 50257, # Vocabulary size
    "context_length": 256, # Context length
    "emb_dim": 768, # Embedding dimension
    "n_heads": 12, # Number of attention heads
    "n_layers": 12, # Number of layers
    "drop_rate": 0.1, # Dropout rate
    "qkv_bias": False # Query-Key-Value bias
}

def generate_text_simple(model, idx,
                        max_new_tokens, context_size):
    # 一次一次的迭代续写
    for _ in range(max_new_tokens):
        idx_cond = idx[:, -context_size:]
        with torch.no_grad():
            logits = model(idx_cond)
            
        logits = logits[:, -1, :]  # The last time step
        probas = torch.softmax(logits, dim=-1)
        idx_next = torch.argmax(probas, dim=-1, keepdim=True)
        idx = torch.cat((idx, idx_next), dim=-1)
    return idx

def text_to_token_ids(text, tokenizer):
    encoded = tokenizer.encode(text, allowed_special={'<|endoftext|>'})
    encoded_tensor = torch.tensor(encoded).unsqueeze(0)
    return encoded_tensor

def token_ids_to_text(token_ids, tokenizer):
    flat = token_ids.squeeze(0)
    return tokenizer.decode(flat.tolist())

损失函数

def calc_loss_batch(input_batch, target_batch, model, device):
    input_batch = input_batch.to(device)
    target_batch = target_batch.to(device)
    logits = model(input_batch)
    loss = torch.nn.functional.cross_entropy(
        logits.flatten(0, 1), target_batch.flatten()
    )
    return loss

def calc_loss_loader(data_loader, model, device, num_batches=None):
    total_loss = 0.
    if len(data_loader) == 0:
        return float("nan")
    elif num_batches is None:
        num_batches = len(data_loader)
    else:
        num_batches = min(num_batches, len(data_loader))
    for i, (input_batch, target_batch) in enumerate(data_loader):
        if i < num_batches:
            loss = calc_loss_batch(
                input_batch, target_batch, model, device 
            )
            total_loss += loss.item()
        else:
            break
    return total_loss / num_batches

预训练函数

def train_model_simple(model, train_loader, val_loader,
                    optimizer, device, num_epochs,
                    eval_freq, eval_iter, start_context, tokenizer):
    train_losses, val_losses, track_tokens_seen = [], [], []
    tokens_seen, global_step = 0, -1
    for epoch in range(num_epochs):
        model.train()
        for input_batch, target_batch in train_loader:
            optimizer.zero_grad()
            loss = calc_loss_batch(
                input_batch, target_batch, model, device
            )
            loss.backward()
            optimizer.step()
            tokens_seen += input_batch.numel()
            global_step += 1
            
            if global_step % eval_freq == 0:
                train_loss, val_loss = evaluate_model(
                    model, train_loader, val_loader, device, eval_iter
                )
                train_losses.append(train_loss)
                val_losses.append(val_loss)
                track_tokens_seen.append(tokens_seen)
                print(f"Ep {epoch+1} (Step {global_step:06d}): "
                    f"Train loss {train_loss:.3f}, "
                    f"Val loss {val_loss:.3f}")
        generate_and_print_sample(
            model, tokenizer, device, start_context
        )
    return train_losses, val_losses, track_tokens_seen

def evaluate_model(model, train_loader, val_loader, device, eval_iter):
    model.eval()
    with torch.no_grad():
        train_loss = calc_loss_loader(
            train_loader, model, device, num_batches=eval_iter
        )
        val_loss = calc_loss_loader(
            val_loader, model, device, num_batches=eval_iter
        )
    model.train()
    return train_loss, val_loss

def generate_and_print_sample(model, tokenizer, device, start_context):
    model.eval()
    context_size = model.pos_emb.weight.shape[0]
    encoded = text_to_token_ids(start_context, tokenizer).to(device)
    with torch.no_grad():
        token_ids = generate_text_simple(
            model=model, idx=encoded,
            max_new_tokens=50, context_size=context_size
        )
    decoded_text = token_ids_to_text(token_ids, tokenizer)
    print(decoded_text.replace("\n", " "))
    model.train()

加载数据类

class GPTDatasetV1(Dataset):
    def __init__(self, txt, tokenizer, max_length, stride):
        self.input_ids = []
        self.target_ids = []
        
        token_ids = tokenizer.encode(txt)
        for i in range(0, len(token_ids) - max_length, stride):
            input_chunk = token_ids[i:i+max_length]
            target_chunk = token_ids[i+1:i+max_length+1]    
            self.input_ids.append(torch.tensor(input_chunk))
            self.target_ids.append(torch.tensor(target_chunk))
    
    def __len__(self):
        return len(self.input_ids)

    def __getitem__(self, idx):
        return self.input_ids[idx], self.target_ids[idx]

def create_dataloader_v1(txt, batch_size=4, max_length=256,
                        stride=128, shuffle=True, drop_last=True,
                        num_workers=0):
    tokenizer = tiktoken.get_encoding("gpt2")
    dataset = GPTDatasetV1(txt, tokenizer, max_length, stride)
    dataloader = DataLoader(
        dataset,
        batch_size=batch_size,
        shuffle=shuffle,
        drop_last=drop_last,
        num_workers=num_workers
    )
    return dataloader

下载预训练数据

import urllib.request
url = ("https://raw.githubusercontent.com/rasbt/"
    "LLMs-from-scratch/main/ch02/01_main-chapter-code/"
    "the-verdict.txt")
file_path = "the-verdict.txt"
urllib.request.urlretrieve(url, file_path)

file_path = "the-verdict.txt"
with open(file_path, "r", encoding="utf-8") as file:
    text_data = file.read()
    
tokenizer = tiktoken.get_encoding("gpt2")
total_characters = len(text_data)
total_tokens = len(tokenizer.encode(text_data))

train_ratio = 0.90
split_idx = int(train_ratio * len(text_data))
train_data = text_data[:split_idx]
val_data = text_data[split_idx:]
torch.manual_seed(123)

train_loader = create_dataloader_v1(
    train_data,
    batch_size=2,
    max_length=GPT_CONFIG_124M["context_length"],
    stride=GPT_CONFIG_124M["context_length"],
    drop_last=True,
    shuffle=True,
    num_workers=0
)

val_loader = create_dataloader_v1(
    val_data,
    batch_size=2,
    max_length=GPT_CONFIG_124M["context_length"],
    stride=GPT_CONFIG_124M["context_length"],
    drop_last=True,
    shuffle=True,
    num_workers=0
)

print("Train loader:")
for x, y in train_loader:
    print(x.shape, y.shape)
    
print("\nValidation loader:")
for x,y in val_loader:
    print(x.shape, y.shape)

Train loader:
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])
torch.Size([2, 256]) torch.Size([2, 256])

Validation loader:
torch.Size([2, 256]) torch.Size([2, 256])

预训练

torch.manual_seed(123)
model = GPTModel(GPT_CONFIG_124M)
# model.eval()

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
with torch.no_grad():
    train_loss = calc_loss_loader(train_loader, model, device)
    val_loss = calc_loss_loader(val_loader, model, device)
    
print("Training loss:", train_loss)
print("Validation loss:", val_loss)

Training loss: 10.995166460673014
Validation loss: 10.99630355834961

optimizer = torch.optim.AdamW(
    model.parameters(),
    lr=0.0004,
    weight_decay=0.1
)
num_epochs = 10 
train_losses, val_losses, tokens_seen = train_model_simple(
    model, train_loader, val_loader, optimizer, device,
    num_epochs=num_epochs, eval_freq=5, eval_iter=5,
    start_context="Every effort moves you", tokenizer=tokenizer
)

Ep 1 (Step 000000): Train loss 9.973, Val loss 9.913
Ep 1 (Step 000005): Train loss 8.197, Val loss 8.290
Every effort moves you                                                  
Ep 2 (Step 000010): Train loss 6.802, Val loss 7.059
Ep 2 (Step 000015): Train loss 6.140, Val loss 6.640
Every effort moves you, the, the, the, the, the, the.                                     
Ep 3 (Step 000020): Train loss 5.688, Val loss 6.512
Ep 3 (Step 000025): Train loss 5.370, Val loss 6.407
Every effort moves you.                                                 
Ep 4 (Step 000030): Train loss 4.856, Val loss 6.393
Ep 4 (Step 000035): Train loss 4.330, Val loss 6.221
Every effort moves you know the, and my the fact--I had the fact of the fact--his, and I had been the fact--I had been, and, and I had been, and--his, and Mrs.       
Ep 5 (Step 000040): Train loss 3.943, Val loss 6.078
Every effort moves you know," was one of the ax.                                          
Ep 6 (Step 000045): Train loss 3.061, Val loss 6.133
Ep 6 (Step 000050): Train loss 2.863, Val loss 6.081
Every effort moves you know," was not that the axioms he was not the the Sevres and silver of an exquisburn, I had to see a smile, and I looked at the donkey again.           
Ep 7 (Step 000055): Train loss 2.209, Val loss 6.160
Ep 7 (Step 000060): Train loss 1.692, Val loss 6.160
Every effort moves you know," was one of the axioms he had been the tips of a self-confident moustache, I felt to see a smile behind his painting.                
Ep 8 (Step 000065): Train loss 1.277, Val loss 6.224
Ep 8 (Step 000070): Train loss 0.978, Val loss 6.216
Every effort moves you?"  "Yes--quite insensible to the fact with the Sevres and silver of an exquisburn's an unusual degree to the display of his close grayish beard--as if he had the donkey. "There were days when I
Ep 9 (Step 000075): Train loss 0.666, Val loss 6.302
Ep 9 (Step 000080): Train loss 0.533, Val loss 6.382
Every effort moves you?"  "Yes--quite insensible to the fact with a laugh: "Yes--and by me!"  "Oh, and Mrs. Stroud. She, one might put it, the donkey. "strongest," as his
Ep 10 (Step 000085): Train loss 0.347, Val loss 6.424
Every effort moves you?"  "Yes--quite insensible to the irony. She wanted him vindicated--and by me!"  He laughed again, and threw back his head to look up at the sketch of the donkey. "There were days when I

matplotlib

import matplotlib.pyplot as plt 
from matplotlib.ticker import MaxNLocator

def plot_losses(epochs_seen, tokens_seen, train_losses, val_losses):
    fig, axl = plt.subplots(figsize=(5,3))
    axl.plot(epochs_seen, train_losses, label="Training loss")
    axl.plot(
        epochs_seen, val_losses, linestyle="-.", label="Validation loss"
    )
    axl.set_xlabel("Epochs")
    axl.set_ylabel("Loss")
    axl.legend(loc="upper right")
    axl.xaxis.set_major_locator(MaxNLocator(integer=True))
    ax2 = axl.twiny()
    ax2.plot(tokens_seen, train_losses, alpha=0)
    ax2.set_xlabel("Tokens seen")
    fig.tight_layout()
    plt.show()
    
epochs_tensor = torch.linspace(0, num_epochs, len(train_losses))
plot_losses(epochs_tensor, tokens_seen, train_losses, val_losses)