File size: 5,496 Bytes
d666c9e |
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 |
import torch
from torch.utils.data import Dataset, DataLoader
from torch import nn
from datasets import load_dataset, concatenate_datasets
from tokenizers import Tokenizer, models, trainers
import math
# --------------------------------------------------
# 1. Loading datasets from Hugging Face
# --------------------------------------------------
def load_hf_datasets():
"""Load and concatenate datasets"""
bookcorpus = load_dataset("bookcorpus", split="train") # 11K books
wiki = load_dataset("wikitext", "wikitext-103-raw-v1", split="train") # Wikipedia
fineweb = load_dataset("fineweb", split="train")
arabic_raw_text = load_dataset("ARABIC-RAW-TEXT", split="train")
tinybooks = load_dataset("tiny-textbooks", split="train")
cc_trajectories = load_dataset("CC-Bench-trajectories", split="train")
textbook = load_dataset("TextbookReasoning", split="train")
megascience = load_dataset("MegaScience", split="train")
return concatenate_datasets([bookcorpus, wiki, fineweb, arabic_raw_text, tinybooks, cc_trajectories, textbook, megascience])
# --------------------------------------------------
# 2. Tokenization (BPE)
# --------------------------------------------------
def train_tokenizer(dataset, vocab_size=30000):
"""Train a Byte-Level BPE tokenizer"""
tokenizer = Tokenizer(models.BPE())
trainer = trainers.BpeTrainer(
vocab_size=vocab_size,
special_tokens=["[PAD]", "[UNK]", "[CLS]", "[SEP]"]
)
# Train on dataset texts
def batch_iterator(batch_size=1000):
for i in range(0, len(dataset), batch_size):
yield dataset[i:i+batch_size]["text"]
tokenizer.train_from_iterator(batch_iterator(), trainer=trainer)
return tokenizer
# --------------------------------------------------
# 3. Preparing DataLoader
# --------------------------------------------------
class TextDataset(Dataset):
def __init__(self, encoded_text, seq_length=128):
self.data = encoded_text
self.seq_length = seq_length
def __len__(self):
return len(self.data) - self.seq_length
def __getitem__(self, idx):
x = self.data[idx:idx+self.seq_length]
y = self.data[idx+1:idx+self.seq_length+1]
return torch.tensor(x), torch.tensor(y)
# --------------------------------------------------
# 4. Transformer Model
# --------------------------------------------------
class TransformerModel(nn.Module):
def __init__(self, vocab_size, d_model=512, nhead=8, num_layers=6):
super().__init__()
self.embedding = nn.Embedding(vocab_size, d_model)
self.pos_encoder = PositionalEncoding(d_model)
encoder_layer = nn.TransformerEncoderLayer(d_model, nhead, dim_feedforward=d_model*4)
self.transformer = nn.TransformerEncoder(encoder_layer, num_layers)
self.fc = nn.Linear(d_model, vocab_size)
def forward(self, x):
x = self.embedding(x) * torch.sqrt(torch.tensor(self.embedding.embedding_dim))
x = self.pos_encoder(x)
x = self.transformer(x)
return self.fc(x)
class PositionalEncoding(nn.Module):
def __init__(self, d_model, max_len=5000):
super().__init__()
pe = torch.zeros(max_len, d_model)
position = torch.arange(0, max_len, dtype=torch.float).unsqueeze(1)
div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))
pe[:, 0::2] = torch.sin(position * div_term)
pe[:, 1::2] = torch.cos(position * div_term)
self.register_buffer('pe', pe)
def forward(self, x):
return x + self.pe[:x.size(1), :]
# --------------------------------------------------
# 5. Training and Generation
# --------------------------------------------------
def main():
# Configuration
SEQ_LENGTH = 128
BATCH_SIZE = 64
VOCAB_SIZE = 30000
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
# 1. Load data
dataset = load_hf_datasets()
# 2. Tokenization
tokenizer = train_tokenizer(dataset, VOCAB_SIZE)
encoded_text = tokenizer.encode(dataset["text"]).ids
# 3. DataLoader
train_dataset = TextDataset(encoded_text, SEQ_LENGTH)
dataloader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True)
# 4. Model
model = TransformerModel(VOCAB_SIZE).to(DEVICE)
optimizer = torch.optim.Adam(model.parameters(), lr=3e-4)
criterion = nn.CrossEntropyLoss()
# 5. Training
for epoch in range(10):
for batch_x, batch_y in dataloader:
batch_x, batch_y = batch_x.to(DEVICE), batch_y.to(DEVICE)
optimizer.zero_grad()
logits = model(batch_x)
loss = criterion(logits.view(-1, VOCAB_SIZE), batch_y.view(-1))
loss.backward()
optimizer.step()
print(f"Epoch {epoch}, Loss: {loss.item():.4f}")
# 6. Text generation
def generate(prompt, max_length=100, temperature=0.7):
model.eval()
tokens = tokenizer.encode(prompt).ids
for _ in range(max_length):
with torch.no_grad():
logits = model(torch.tensor([tokens[-SEQ_LENGTH:]]).to(DEVICE))
probs = torch.softmax(logits[0, -1] / temperature, dim=-1)
next_token = torch.multinomial(probs, num_samples=1).item()
tokens.append(next_token)
return tokenizer.decode(tokens)
print(generate("The meaning of life is"))
if __name__ == "__main__":
main() |