|
Supplementary code for the Build a Large Language Model From Scratch book by Sebastian Raschka Code repository: https://github.com/rasbt/LLMs-from-scratch |
|
Multi-head Attention Plus Data Loading#
# NBVAL_IGNORE_OUTPUT
from importlib.metadata import version
print("torch version:", version("torch"))
---------------------------------------------------------------------------
PackageNotFoundError Traceback (most recent call last)
Cell In[1], line 4
1 # NBVAL_IGNORE_OUTPUT
2 from importlib.metadata import version
----> 4 print("torch version:", version("torch"))
File /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/importlib/metadata/__init__.py:946, in version(distribution_name)
939 def version(distribution_name):
940 """Get the version string for the named package.
941
942 :param distribution_name: The name of the distribution package to query.
943 :return: The version string for the package as defined in the package's
944 "Version" metadata key.
945 """
--> 946 return distribution(distribution_name).version
File /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/importlib/metadata/__init__.py:919, in distribution(distribution_name)
913 def distribution(distribution_name):
914 """Get the ``Distribution`` instance for the named package.
915
916 :param distribution_name: The name of the distribution package as a string.
917 :return: A ``Distribution`` instance (or subclass thereof).
918 """
--> 919 return Distribution.from_name(distribution_name)
File /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/importlib/metadata/__init__.py:518, in Distribution.from_name(cls, name)
516 return dist
517 else:
--> 518 raise PackageNotFoundError(name)
PackageNotFoundError: No package metadata was found for torch
The complete chapter code is located in ch03.ipynb.
This notebook contains the main takeaway, multihead-attention implementation (plus the data loading pipeline from chapter 2)
Data Loader from Chapter 2#
import tiktoken
import torch
import torch.nn as nn
from torch.utils.data import Dataset, DataLoader
class GPTDatasetV1(Dataset):
def __init__(self, txt, tokenizer, max_length, stride):
self.input_ids = []
self.target_ids = []
# Tokenize the entire text
token_ids = tokenizer.encode(txt, allowed_special={'<|endoftext|>'})
# Use a sliding window to chunk the book into overlapping sequences of max_length
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(txt, batch_size=4, max_length=256, stride=128, shuffle=True):
# Initialize the tokenizer
tokenizer = tiktoken.get_encoding("gpt2")
# Create dataset
dataset = GPTDatasetV1(txt, tokenizer, max_length, stride)
# Create dataloader
dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=shuffle)
return dataloader
with open("small-text-sample.txt", "r", encoding="utf-8") as f:
raw_text = f.read()
tokenizer = tiktoken.get_encoding("gpt2")
encoded_text = tokenizer.encode(raw_text)
vocab_size = 50257
output_dim = 256
max_len = 1024
context_length = max_len
token_embedding_layer = nn.Embedding(vocab_size, output_dim)
pos_embedding_layer = torch.nn.Embedding(context_length, output_dim)
max_length = 4
dataloader = create_dataloader(raw_text, batch_size=8, max_length=max_length, stride=max_length)
for batch in dataloader:
x, y = batch
token_embeddings = token_embedding_layer(x)
pos_embeddings = pos_embedding_layer(torch.arange(max_length))
input_embeddings = token_embeddings + pos_embeddings
break
print(input_embeddings.shape)
torch.Size([8, 4, 256])
Multi-head Attention from Chapter 3#
Variant A: Simple implementation#
class CausalSelfAttention(nn.Module):
def __init__(self, d_in, d_out, context_length, dropout, qkv_bias=False):
super().__init__()
self.d_out = d_out
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.dropout = nn.Dropout(dropout) # New
self.register_buffer('mask', torch.triu(torch.ones(context_length, context_length), diagonal=1)) # New
def forward(self, x):
b, n_tokens, d_in = x.shape # New batch dimension b
keys = self.W_key(x)
queries = self.W_query(x)
values = self.W_value(x)
attn_scores = queries @ keys.transpose(1, 2) # Changed transpose
attn_scores.masked_fill_( # New, _ ops are in-place
self.mask.bool()[:n_tokens, :n_tokens], -torch.inf)
attn_weights = torch.softmax(attn_scores / keys.shape[-1]**0.5, dim=-1)
attn_weights = self.dropout(attn_weights) # New
context_vec = attn_weights @ values
return context_vec
class MultiHeadAttentionWrapper(nn.Module):
def __init__(self, d_in, d_out, context_length, dropout, num_heads, qkv_bias=False):
super().__init__()
self.heads = nn.ModuleList(
[CausalSelfAttention(d_in, d_out, context_length, dropout, qkv_bias)
for _ in range(num_heads)]
)
self.out_proj = nn.Linear(d_out*num_heads, d_out*num_heads)
def forward(self, x):
context_vec = torch.cat([head(x) for head in self.heads], dim=-1)
return self.out_proj(context_vec)
torch.manual_seed(123)
context_length = max_length
d_in = output_dim
num_heads=2
d_out = d_in // num_heads
mha = MultiHeadAttentionWrapper(d_in, d_out, context_length, 0.0, num_heads)
batch = input_embeddings
context_vecs = mha(batch)
print("context_vecs.shape:", context_vecs.shape)
context_vecs.shape: torch.Size([8, 4, 256])
Variant B: Alternative implementation#
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 be divisible by num_heads"
self.d_out = d_out
self.num_heads = num_heads
self.head_dim = d_out // num_heads # Reduce the projection dim to match desired output dim
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) # Linear layer to combine head outputs
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
keys = self.W_key(x) # Shape: (b, num_tokens, d_out)
queries = self.W_query(x)
values = self.W_value(x)
# We implicitly split the matrix by adding a `num_heads` dimension
# Unroll last dim: (b, num_tokens, d_out) -> (b, num_tokens, num_heads, head_dim)
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)
# Transpose: (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)
# Compute scaled dot-product attention (aka self-attention) with a causal mask
attn_scores = queries @ keys.transpose(2, 3) # Dot product for each head
# Original mask truncated to the number of tokens and converted to boolean
mask_bool = self.mask.bool()[:num_tokens, :num_tokens]
# Use the mask to fill attention scores
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)
# Shape: (b, num_tokens, num_heads, head_dim)
context_vec = (attn_weights @ values).transpose(1, 2)
# Combine heads, where self.d_out = self.num_heads * self.head_dim
context_vec = context_vec.contiguous().view(b, num_tokens, self.d_out)
context_vec = self.out_proj(context_vec) # optional projection
return context_vec
torch.manual_seed(123)
context_length = max_length
d_in = output_dim
d_out = d_in
mha = MultiHeadAttention(d_in, d_out, context_length, 0.0, num_heads=2)
batch = input_embeddings
context_vecs = mha(batch)
print("context_vecs.shape:", context_vecs.shape)
context_vecs.shape: torch.Size([8, 4, 256])