-
Notifications
You must be signed in to change notification settings - Fork 0
/
architecture.py
159 lines (126 loc) · 5.1 KB
/
architecture.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
import torch
import torch.nn as nn
from einops import rearrange
class GPTModel(nn.Module):
def __init__(self, cfg):
super().__init__()
# Token and position embeddings
self.token_emb = nn.Embedding(cfg["vocab_size"], cfg["emb_dim"])
self.position_emb = nn.Embedding(cfg["context_length"], cfg["emb_dim"])
self.dropout = nn.Dropout(cfg["dropout"])
# Transformer blocks
self.trf_blocks = nn.Sequential(
*[TransformerBlock(cfg) for _ in range(cfg["n_layers"])]
)
# Final normalization layer and output projection
self.final_norm = LayerNorm(cfg["emb_dim"])
self.out_proj = nn.Linear(cfg["emb_dim"], cfg["vocab_size"], bias=False)
def forward(self, x):
batch_size, num_tokens = x.shape
token_embeds = self.token_emb(x)
position_embeds = self.position_emb(torch.arange(num_tokens).to(x.device))
# Add token and position embeddings
# x shape: [batch_size, num_tokens, emb_dim]
x = token_embeds + position_embeds
x = self.dropout(x)
x = self.trf_blocks(x)
x = self.final_norm(x)
x = self.out_proj(x)
return x
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
class GELU(nn.Module):
def __init__(self):
super().__init__()
def forward(self, x):
# Gaussian Error Linear Unit activation
out = 0.5 * x * (1 + torch.tanh(
torch.sqrt(torch.tensor(2.0 / torch.pi)) *
(x + 0.044715 * torch.pow(x, 3))
))
return out
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)
class MultiHeadCausalAttention(nn.Module):
# d_in: input token embed dim, d_out: q, k, and v space dim
def __init__(self, d_in, d_out, sample_length, num_heads, dropout, qkv_bias=False):
super().__init__()
assert (d_out % num_heads == 0), \
"d_out must be divisible by num_heads"
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.dropout = nn.Dropout(dropout)
self.register_buffer('mask',
torch.triu(torch.ones(sample_length, sample_length), diagonal=1))
self.out_proj = nn.Linear(d_out, d_out)
def forward(self, x):
# x shape: [batch_size, num_tokens, d_in]
batch_size, num_tokens, d_in = x.shape
# q, k, and v shape: [batch_size, num_tokens, d_out]
queries = self.W_query(x)
keys = self.W_key(x)
values = self.W_value(x)
# Split the matrix by adding a "num_heads" dimension
# q, k, and v shape: [batch_size, num_heads, num_tokens, head_dim]
queries, keys, values = map(
lambda x: rearrange(x, 'b s (h d) -> b h s d', h=self.num_heads),
[queries, keys, values]
)
attn_scores = queries @ keys.transpose(2, 3)
# Set mask size according to the number of tokens
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)
attention = attn_weights @ values
# attention shape: [batch_size, num_tokens, d_out]
attention = rearrange(attention, 'b h s d -> b s (h d)')
attention = self.out_proj(attention)
return attention
class TransformerBlock(nn.Module):
def __init__(self, cfg):
super().__init__()
self.msa = MultiHeadCausalAttention(
d_in=cfg["emb_dim"],
d_out=cfg["emb_dim"],
sample_length=cfg["context_length"],
num_heads=cfg["n_heads"],
dropout=cfg["dropout"],
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["dropout"])
def forward(self, x):
shortcut = x
x = self.norm1(x)
x = self.msa(x)
x = self.drop_shortcut(x)
x = shortcut + x
shortcut = x
x = self.norm2(x)
x = self.ff(x)
x = self.drop_shortcut(x)
out = shortcut + x
return out