Spaces:
Sleeping
Sleeping
File size: 11,061 Bytes
6442065 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 |
import os
import math
import time
import torch
import torch.nn as nn
from torch.nn import functional as F
from dataclasses import dataclass
from torch.nn.parallel import DistributedDataParallel as DDP
import numpy as np
from datetime import datetime
# Hyperparameters
learning_rate = 3e-4 # Peak learning rate
min_lr = 3e-5 # Minimum learning rate at the end of training
warmup_iters = 2000 # Linear warmup over warmup_iters
lr_decay_iters = 800000 # Cosine decay over lr_decay_iters
weight_decay = 0.1 # AdamW weight decay
beta1 = 0.9 # AdamW beta1
beta2 = 0.95 # AdamW beta2
grad_clip = 1.0 # Clip gradients at this value
decay_lr = True # Whether to decay learning rate
batch_size = 64 # Training batch size
block_size = 256 # Maximum sequence length
eval_interval = 500 # How often to evaluate
eval_iters = 200 # Number of iterations to use for evaluation
log_interval = 10 # How often to print training progress
# Model architecture
@dataclass
class GPTConfig:
block_size: int = block_size
vocab_size: int = 50304
n_layer: int = 12
n_head: int = 16
n_embd: int = 1024
dropout: float = 0.1
bias: bool = False
class CausalSelfAttention(nn.Module):
def __init__(self, config):
super().__init__()
assert config.n_embd % config.n_head == 0
self.c_attn = nn.Linear(config.n_embd, 3 * config.n_embd, bias=config.bias)
self.c_proj = nn.Linear(config.n_embd, config.n_embd, bias=config.bias)
self.attn_dropout = nn.Dropout(config.dropout)
self.resid_dropout = nn.Dropout(config.dropout)
self.n_head = config.n_head
self.n_embd = config.n_embd
self.dropout = config.dropout
def forward(self, x):
B, T, C = x.size()
qkv = self.c_attn(x)
q, k, v = qkv.split(self.n_embd, dim=2)
k = k.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)
q = q.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)
v = v.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)
y = F.scaled_dot_product_attention(q, k, v, is_causal=True)
y = y.transpose(1, 2).contiguous().view(B, T, C)
y = self.resid_dropout(self.c_proj(y))
return y
class MLP(nn.Module):
def __init__(self, config):
super().__init__()
self.c_fc = nn.Linear(config.n_embd, 4 * config.n_embd, bias=config.bias)
self.gelu = nn.GELU()
self.c_proj = nn.Linear(4 * config.n_embd, config.n_embd, bias=config.bias)
self.dropout = nn.Dropout(config.dropout)
def forward(self, x):
x = self.c_fc(x)
x = self.gelu(x)
x = self.c_proj(x)
x = self.dropout(x)
return x
class Block(nn.Module):
def __init__(self, config):
super().__init__()
self.ln_1 = nn.LayerNorm(config.n_embd)
self.attn = CausalSelfAttention(config)
self.ln_2 = nn.LayerNorm(config.n_embd)
self.mlp = MLP(config)
def forward(self, x):
x = x + self.attn(self.ln_1(x))
x = x + self.mlp(self.ln_2(x))
return x
class GPT(nn.Module):
def __init__(self, config):
super().__init__()
self.config = config
self.transformer = nn.ModuleDict(dict(
wte = nn.Embedding(config.vocab_size, config.n_embd),
wpe = nn.Embedding(config.block_size, config.n_embd),
drop = nn.Dropout(config.dropout),
h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
ln_f = nn.LayerNorm(config.n_embd)
))
self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
self.transformer.wte.weight = self.lm_head.weight
# Initialize weights
self.apply(self._init_weights)
for pn, p in self.named_parameters():
if pn.endswith('c_proj.weight'):
torch.nn.init.normal_(p, mean=0.0, std=0.02/math.sqrt(2 * config.n_layer))
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, idx, targets=None):
device = idx.device
b, t = idx.size()
pos = torch.arange(0, t, dtype=torch.long, device=device)
tok_emb = self.transformer.wte(idx)
pos_emb = self.transformer.wpe(pos)
x = self.transformer.drop(tok_emb + pos_emb)
for block in self.transformer.h:
x = block(x)
x = self.transformer.ln_f(x)
if targets is not None:
logits = self.lm_head(x)
loss = F.cross_entropy(logits.view(-1, logits.size(-1)), targets.view(-1), ignore_index=-1)
else:
logits = self.lm_head(x[:, [-1], :])
loss = None
return logits, loss
@torch.no_grad()
def generate(self, idx, max_new_tokens, temperature=1.0, top_k=None):
for _ in range(max_new_tokens):
idx_cond = idx if idx.size(1) <= self.config.block_size else idx[:, -self.config.block_size:]
logits, _ = self(idx_cond)
logits = logits[:, -1, :] / temperature
if top_k is not None:
v, _ = torch.topk(logits, min(top_k, logits.size(-1)))
logits[logits < v[:, [-1]]] = -float('Inf')
probs = F.softmax(logits, dim=-1)
idx_next = torch.multinomial(probs, num_samples=1)
idx = torch.cat((idx, idx_next), dim=1)
return idx
def get_batch(data, block_size, batch_size):
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+1+block_size] for i in ix])
return x, y
def get_lr(it):
# 1) Linear warmup for warmup_iters steps
if it < warmup_iters:
return learning_rate * it / warmup_iters
# 2) If it > lr_decay_iters, return min learning rate
if it > lr_decay_iters:
return min_lr
# 3) In between, use cosine decay down to min learning rate
decay_ratio = (it - warmup_iters) / (lr_decay_iters - warmup_iters)
assert 0 <= decay_ratio <= 1
coeff = 0.5 * (1.0 + math.cos(math.pi * decay_ratio))
return min_lr + coeff * (learning_rate - min_lr)
def save_training_log(log_entry, filename='training_logs.md'):
"""Save training logs in markdown format"""
timestamp = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
with open(filename, 'a') as f:
if not f.tell(): # If file is empty, write header
f.write('# Training Logs\n\n')
f.write('| Timestamp | Iteration | Training Loss | Learning Rate |\n')
f.write('|-----------|------------|---------------|---------------|\n')
f.write(f'| {timestamp} | {log_entry["iter"]:10d} | {log_entry["train_loss"]:.6f} | {log_entry["lr"]:.2e} |\n')
def save_model(model, optimizer, iter_num, loss, filename):
"""Save model checkpoint with error handling"""
try:
# First save to a temporary file
tmp_filename = filename + '.tmp'
checkpoint = {
'model_state_dict': model.state_dict(),
'optimizer_state_dict': optimizer.state_dict(),
'iter_num': iter_num,
'loss': loss,
}
# Use torch.save with zip compression
torch.save(checkpoint, tmp_filename, _use_new_zipfile_serialization=True)
# If save was successful, rename tmp file to final filename
if os.path.exists(filename):
os.remove(filename) # Remove old file if it exists
os.rename(tmp_filename, filename)
return True
except Exception as e:
print(f"Error saving model to {filename}: {str(e)}")
# Clean up temp file if it exists
if os.path.exists(tmp_filename):
try:
os.remove(tmp_filename)
except:
pass
return False
def main():
torch.manual_seed(1337)
torch.backends.cuda.matmul.allow_tf32 = True
torch.backends.cudnn.allow_tf32 = True
device = 'cuda' if torch.cuda.is_available() else 'cpu'
print(f"Using device: {device}")
# Create checkpoint directory
os.makedirs('checkpoints', exist_ok=True)
# Load the data
with open('input.txt', 'r') as f:
text = f.read()
chars = sorted(list(set(text)))
vocab_size = len(chars)
stoi = {ch:i for i,ch in enumerate(chars)}
itos = {i:ch for i,ch in enumerate(chars)}
encode = lambda s: [stoi[c] for c in s]
data = torch.tensor(encode(text), dtype=torch.long)
n = int(0.9 * len(data))
train_data = data[:n]
val_data = data[n:]
# Initialize the model
model = GPT(GPTConfig(vocab_size=vocab_size))
model = model.to(device)
print(f"Model parameters: {sum(p.numel() for p in model.parameters())/1e6:.2f}M")
# Initialize optimizer
optimizer = torch.optim.AdamW(model.parameters(), lr=learning_rate, betas=(beta1, beta2), weight_decay=weight_decay)
# Training loop
best_train_loss = float('inf')
iter_num = 0
while True:
# Get batch and learning rate
xb, yb = get_batch(train_data, block_size, batch_size)
xb, yb = xb.to(device), yb.to(device)
lr = get_lr(iter_num) if decay_lr else learning_rate
for param_group in optimizer.param_groups:
param_group['lr'] = lr
# Forward pass
logits, loss = model(xb, yb)
optimizer.zero_grad(set_to_none=True)
loss.backward()
torch.nn.utils.clip_grad_norm_(model.parameters(), grad_clip)
optimizer.step()
# Logging and model saving
if iter_num % log_interval == 0:
train_loss = loss.item()
print(f"iter {iter_num}: loss {train_loss:.4f}, lr {lr:e}")
save_training_log({
"iter": iter_num,
"train_loss": train_loss,
"lr": lr
})
# Save model if loss improved
if train_loss < best_train_loss:
best_train_loss = train_loss
print(f"Saving model with training loss: {best_train_loss:.6f}")
# Save the latest model
save_model(
model,
optimizer,
iter_num,
best_train_loss,
os.path.join('checkpoints', 'latest_model.pt')
)
if best_train_loss < 0.099999:
print(f"Achieved target loss of {best_train_loss:.6f}")
break
iter_num += 1
if iter_num > lr_decay_iters:
break
if __name__ == '__main__':
main() |