-
Notifications
You must be signed in to change notification settings - Fork 0
/
simple_lm.py
111 lines (93 loc) · 3.87 KB
/
simple_lm.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
import torch
import torch.nn as nn
import torch.nn.functional as F
# hyperparameters
batch_size = 32
block_size = 8 # context length
max_iters = 10000
eval_interval = 1000
learning_rate = 0.001
device = 'cude' if torch.cuda.is_available() else 'cpu'
eval_iters = 200
# --------------------------------------------
torch.manual_seed(1337)
with open('input.txt', 'r', encoding='utf-8') as f: text = f.read()
# all the unique characters
chars = sorted(list(set(text)))
vocab_size = len(chars)
# char to index and index to char
char_to_idx = {ch: i for i, ch in enumerate(chars)}
idx_to_char = {i: ch for i, ch in enumerate(chars)}
encode = lambda s: [char_to_idx[c] for c in s]
decode = lambda li: ''.join([idx_to_char[i] for i in li])
# train-test split
data = torch.tensor(encode(text), dtype=torch.long, device=device)
split = int(0.9 * len(data))
train_data, val_data = data[:split], data[split:]
# data loader
def get_batch(split):
data = train_data if split == 'train' else val_data
ix = torch.randint(high=len(data) - batch_size, size=(batch_size,)) # random indices betweeen 0 to len(data)-batch_size
x = torch.stack([data[i:i+block_size] for i in ix]) # next block_size characters from each random integer
y = torch.stack([data[i+1:i+block_size+1] for i in ix]) # next block_size characters offset by 1
x,y = x.to(device), y.to(device)
return x,y
@torch.no_grad()
def estimate_loss(model):
out = {}
model.eval()
for split in ['train', 'val']:
losses = torch.zeros(eval_iters, device=device)
for k in range(eval_iters):
X,Y = get_batch(split)
logits, loss = model(X,Y)
losses[k] = loss.item()
out[split] = losses.mean()
model.train()
return out
class SimpleLanguageModel(nn.Module):
def __init__(self, vocab_size):
super().__init__()
self.token_embedding_table = nn.Embedding(vocab_size, vocab_size) # representing each token as a vector of dim=65 (vocab_size)
def forward(self, idx, targets=None):
# idx and targets are of shape (batch_size, block_size) or (B, T)
logits = self.token_embedding_table(idx) # shape: (batch_size, block_size, vocab_size) or (B, T, C)
if targets is None: loss = None
else:
B, T, C = logits.shape
logits = logits.view(B*T, C)
targets = targets.view(B*T)
loss = F.cross_entropy(logits, targets)
return logits, loss
def generate(self, idx, max_new_tokens):
# idx is (B, T) array of indices in the current context
for _ in range(max_new_tokens):
# get the predictions
logits, loss = self(idx)
# focus on the last time step
logits = logits[:, -1, :] # becomes (B, C)
# apply softmax
probs = F.softmax(logits, dim=-1) # (B, C)
# sample from the distribution
idx_next = torch.multinomial(probs, num_samples=1) # (B, 1)
# append sampled index to the running sequence
idx = torch.cat((idx, idx_next), dim=1) # (B, T+1)
return idx
model = SimpleLanguageModel(vocab_size).to(device)
optimizer = torch.optim.AdamW(model.parameters(), lr=learning_rate)
for iter in range(max_iters):
# estimate loss on train and val datasets
if iter % eval_interval == 0:
losses = estimate_loss(model)
print(f'{iter:5d}/{max_iters}: Train loss - {losses["train"]:.4f} - Val loss - {losses["val"]:.5f}')
# sample a batch of data
xb, yb = get_batch('train')
# evaluate the loss
logits, loss = model(xb, yb)
optimizer.zero_grad(set_to_none=True)
loss.backward()
optimizer.step()
# generate some text
context = torch.tensor(encode('The meaning of life is')).reshape(1, -1).to(device)
generated = model.generate(context, 500)[0].tolist()
print(f"\n[Generated Text]:\n {decode(generated)}")