-
Notifications
You must be signed in to change notification settings - Fork 0
/
trainig.py
255 lines (206 loc) · 8.55 KB
/
trainig.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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
import torch
import torch.nn as nn
import torch.nn.functional as F
import mmap
import random
import pickle
import argparse
parser = argparse.ArgumentParser(description="This is a demonstration program")
# Here we add an argument to the parser, specifying the expected type, a help message etc.
parser.add_argument('-batch_size', type=str, required=True, help='please provide a batch_size')
args = parser.parse_args()
# Now we can use the argument value in our program
print(f"batch size: {args.batch_size}")
device = 'cuda' if torch.cuda.is_available() else 'CPU'
print(device)
block_size = int(args.batch_size)
batch_size = 128
max_iters = 200
learning_rate = 3e-4
eval_iters = 100
n_embd = 384
n_layer = 1
n_head = 1
dropout = 0.2
chars = ""
with open('vocab.txt', 'r', encoding = 'utf-8') as f:
text = f.read()
chars = sorted(set(text))
vocab_size = len(chars)
string_to_int = {ch:i for i,ch in enumerate(chars)}
int_to_string = {i:ch for i,ch in enumerate(chars)}
encode = lambda s: [string_to_int[c] for c in s]
decode = lambda l: ''.join([int_to_string[i] for i in l])
data = torch.tensor(encode(text), dtype = torch.long)
# print(data[:100])
print(data)
def get_random_chunk(split):
filename = "train_split.txt" if split == 'train' else 'val_split.txt'
with open(filename, 'rb') as f:
with mmap.mmap(f.fileno(), 0, access = mmap.ACCESS_READ) as mm:
# Determine a file size and random position to start
file_size = len(mm)
start_pos = random.randint(0, (file_size) - block_size*batch_size)
# Seek to the random position and read the block of text
mm.seek(start_pos)
block = mm.read(block_size*batch_size-1)
# Decode the block to the string, ignoring my invalid byte sequences
decoded_block = block.decode('utf-8', errors='ignore').replace('\r', '')
# Train and test splits
data = torch.tensor(encode(decoded_block), dtype=torch.long)
return data
def get_batch(split):
data = get_random_chunk(split)
ix = torch.randint(len(data) - block_size, (batch_size, ))
x = torch.stack([data[i:i+block_size]for i in ix])
y = torch.stack([data[i+1:i+block_size+1] for i in ix])
x, y = x.to(device), y.to(device)
return x, y
@torch.no_grad()
def estimate_loss():
out = {}
model.eval()
for split in ['train', 'val']:
losses = torch.zeros(eval_iters)
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 Head(nn.Module):
""" one head of self-attention """
def __init__(self, head_size):
super().__init__()
self.key = nn.Linear(n_embd, head_size, bias = False)
self.query = nn.Linear(n_embd, head_size, bias = False)
self.value = nn.Linear(n_embd, head_size, bias = False)
self.register_buffer('tril', torch.tril(torch.ones(block_size, block_size)))
self.dropout = nn.Dropout(dropout)
def forward(self, x):
# input of size (batch, time-step, channels)
# output of size (batch, time-step, head size)
B,T,C = x.shape
k = self.key(x) # (B, T, hs)
q = self.query(x) # (B, T, hs)
# compute attention scores ("affinities")
wei = q @ k.transpose(-2,-1) * k.shape[-1]**-0.5 # (B, T, hs) @ (B, hs, T) -> (B, T, T)
wei = wei.masked_fill(self.tril[:T, :T] == 0, float('-inf')) # (B, T, T)
wei = F.softmax(wei, dim = -1) # (B, T, T)
wei = self.dropout(wei)
# perform the weighted aggregation of the values
v = self.value(x) # (B, T, hs)
out = wei @ v # (B, T, T) @ (B, T, hs) -> (B, T, hs)
return out
class MultiHeadAttention(nn.Module):
""" multiple head of self-attention in parallel """
def __init__(self, num_heads, head_size):
super().__init__()
self.heads = nn.ModuleList([Head(head_size) for _ in range(num_heads)])
self.proj = nn.Linear(head_size * num_heads, n_embd)
self.dropout = nn.Dropout(dropout)
def forward(self, x):
out = torch.cat([h(x) for h in self.heads], dim = - 1)
out = self.dropout(self.proj(out))
return out
class FeedForward(nn.Module):
""" a simple linear layer followed by the non-linearity """
def __init__(self, n_embd):
super().__init__()
self.net = nn.Sequential(
nn.Linear(n_embd, 4 * n_embd),
nn.ReLU(),
nn.Linear(4 * n_embd, n_embd),
nn.Dropout(dropout),
)
def forward(self, x):
return self.net(x)
class Block(nn.Module):
"""
Transformer Block: communication followed by computation
"""
def __init__(self, n_embd, n_head):
# n_embd: embedding dimension, n_head: the number of heads we'd like
super().__init__()
head_size = n_embd // n_head
self.sa = MultiHeadAttention(n_head, head_size) # sa == self-attention
self.ffwd = FeedForward(n_embd)
self.ln1 = nn.LayerNorm(n_embd)
self.ln2 = nn.LayerNorm(n_embd)
def forward(self, x):
y = self.sa(x)
x = self.ln1(x + y)
y = self.ffwd(x)
x = self.ln2(x + y)
return x
class GPTLanguageModel(nn.Module):
def __init__(self, vocab_size):
super().__init__()
self.token_embedding_table = nn.Embedding(vocab_size, n_embd)
self.position_embedding_table = nn.Embedding(block_size, n_embd)
self.blocks = nn.Sequential(*[Block(n_embd, n_head=n_head) for _ in range(n_layer)])
self.ln_f = nn.LayerNorm(n_embd) # final layer norm
self.lm_head = nn.Linear(n_embd, vocab_size)
self.apply(self.__init_weights)
def __init_weights(self, module):
if isinstance(module, nn.Linear):
torch.nn.init.normal_(module.weight, mean=0.0, std = 0.02)
if module.bias is not None:
torch.nn.init.zeros_(module.bias)
elif isinstance(module, nn.Embedding):
torch.nn.init.normal_(module.weight, mean=0.0, std = 0.02)
def forward(self, index, targets=None):
B, T = index.shape
# idx and targets are both (B, T) tensors of integers.
tok_emb = self.token_embedding_table(index) # (B, T, C)
pos_emb = self.position_embedding_table(torch.arange(T, device=device)) # (T, C)
x = tok_emb + pos_emb # (B, T, C)
x = self.blocks(x) # (B, T, C)
x = self.ln_f(x) # (B, T, C)
logits = self.lm_head(x) # (B, T, vocab_size)
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, index, max_new_tokens):
# index is B, T indices in the current context
for _ in range(max_new_tokens):
# get the predictions
logits, loss = self.forward(index)
# focus only on the last time step
logits = logits[:, -1, :] # becomes B, C
# apply softmax to get probabilities
probs = F.softmax(logits, dim=-1) # B, C
# sample from the distribution
index_next = torch.multinomial(probs, num_samples = 1) # B, 1
# append sample index to the running sequence
index = torch.cat((index, index_next), dim = 1) # (B, T+1)
return index
model = GPTLanguageModel(vocab_size)
#print("loding model parameters....")
#with open('model-01.pkl', 'rb') as f:
# model = pickle.load(f)
# print("model loaded succesfully!!")
m = model.to(device)
# Create a pytorch optimizer
optimizer = torch.optim.AdamW(model.parameters(), lr = learning_rate)
for iter in range(max_iters):
if iter % eval_iters == 0:
losses = estimate_loss()
print(f'step: {iter}, train loss: {losses['train']:.4f}, val loss: {losses['val']:.4f}')
# sample a batch of data
xb, yb = get_batch('train')
# evaluate the loss
logits, loss = model.forward(xb, yb)
optimizer.zero_grad(set_to_none=True)
loss.backward()
optimizer.step()
print(loss.item())
with open('model-01.pkl', 'wb') as f:
pickle.dump(model, f)
print('model saved')