parameterized_model.py

import torch
import torch.nn as nn
from transformers import PreTrainedModel
from transformers.modeling_outputs import CausalLMOutput
from torch.nn import functional as F
from transformers import PreTrainedConfig

# hyperparameters
batch_size = 16 # how many independent sequences will we process in parallel?
block_size = 32 # what is the maximum context length for predictions?
max_iters = 5000
eval_interval = 100
learning_rate = 1e-3
device = 'cuda' if torch.cuda.is_available() else 'cpu'
eval_iters = 200
n_embd = 64
n_head = 4
n_layer = 4
dropout = 0.0
# ------------

torch.manual_seed(1337)

# wget https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt
with open('input.txt', 'r', encoding='utf-8') as f:
    text = f.read()

# here are all the unique characters that occur in this text
chars = sorted(list(set(text)))
vocab_size = len(chars)
# create a mapping from characters to integers
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] # encoder: take a string, output a list of integers
decode = lambda l: ''.join([itos[i] for i in l]) # decoder: take a list of integers, output a string

# Train and test splits
data = torch.tensor(encode(text), dtype=torch.long)
n = int(0.9*len(data)) # first 90% will be train, rest val
train_data = data[:n]
val_data = data[n:]

# data loading
def get_batch(split):
    # generate a small batch of data of inputs x and targets y
    data = train_data if split == 'train' else val_data
    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(model, prompt="ROMEO:", max_new_tokens=120):
    out = {}
    model.eval()
    for split in ['train', 'val']:
        losses = torch.zeros(eval_iters)
        for k in range(eval_iters):
            X, Y = get_batch(split)
            outputs = model(
                input_ids=X,
                labels=Y
            )
            losses[k] = outputs.loss.item()
        out[split] = losses.mean()

    # generation sample
    input_ids = torch.tensor(
        [[stoi[c] for c in prompt]],
        dtype=torch.long,
        device=device
    )

    generated = model.generate_custom(
        input_ids,
        max_new_tokens=max_new_tokens
    )

    out["sample_text"] = decode(
        generated[0].tolist()
    )
    model.train()
    return out

class ShakespeareConfig(PreTrainedConfig):
    model_type = "tiny_shakespeare"

    def __init__(
        self,
        vocab_size=65,
        block_size=32,
        n_embd=64,
        n_head=4,
        n_layer=4,
        dropout=0.0,
        **kwargs,
    ):
        super().__init__(**kwargs)

        self.vocab_size = vocab_size
        self.block_size = block_size
        self.n_embd = n_embd
        self.n_head = n_head
        self.n_layer = n_layer
        self.dropout = dropout

class SingleHead(nn.Module):
    """ one head of self-attention """

    def __init__(
        self,
        n_embd,
        head_size,
        block_size,
        dropout
    ):
        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):
        B,T,C = x.shape
        k = self.key(x)   # (B,T,head_size)
        q = self.query(x) # (B,T,head_size)
        # compute attention scores ("affinities")
        wei = q @ k.transpose(-2,-1) * C**-0.5 # (B, T, C) @ (B, C, 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,C)
        out = wei @ v # (B, T, T) @ (B, T, C) -> (B, T, C)
        return out

class MultiHeadAttention(nn.Module):
    """ multiple heads of self-attention in parallel """

    def __init__(
        self,
        n_embd,
        num_heads,
        head_size,
        block_size,
        dropout
    ):
        super().__init__()
        self.heads = nn.ModuleList([
            SingleHead(
                n_embd,
                head_size,
                block_size,
                dropout
            )
            for _ in range(num_heads)
        ])
        self.proj = nn.Linear(n_embd, 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 FeedFoward(nn.Module):
    """ a simple linear layer followed by a non-linearity """

    def __init__(self, n_embd, dropout):
        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,
        block_size,
        dropout
    ):
        # 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_embd,
            n_head,
            head_size,
            block_size,
            dropout
        )
        self.ffwd = FeedFoward(
            n_embd,
            dropout
        )
        self.ln1 = nn.LayerNorm(n_embd)
        self.ln2 = nn.LayerNorm(n_embd)

    def forward(self, x):
        x = x + self.sa(self.ln1(x))
        x = x + self.ffwd(self.ln2(x))
        return x

# super simple bigram model
class BigramLanguageModel(PreTrainedModel):
    config_class = ShakespeareConfig
    def __init__(self, config):
        super().__init__(config)
        # each token directly reads off the logits for the next token from a lookup table
        self.token_embedding_table = nn.Embedding(config.vocab_size, config.n_embd)
        self.position_embedding_table = nn.Embedding(config.block_size, config.n_embd)
        self.blocks = nn.Sequential(*[Block(config.n_embd, n_head=config.n_head, block_size=config.block_size, dropout=config.dropout) for _ in range(config.n_layer)])
        self.ln_f = nn.LayerNorm(config.n_embd) # final layer norm
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size)
        self.post_init()

    def forward(self, input_ids, labels=None, **kwargs):
        B, T = input_ids.shape

        # input_ids and targets are both (B,T) tensor of integers
        tok_emb = self.token_embedding_table(input_ids) # (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 labels is None:
            loss = None
        else:
            B, T, C = logits.shape
            logits = logits.view(B*T, C)
            labels = labels.view(B*T)
            loss = F.cross_entropy(logits, labels)

        return CausalLMOutput(
            loss=loss,
            logits=logits
        )

    def generate_custom(self, input_ids, max_new_tokens):
        # idx is (B, T) array of indices in the current context
        idx = input_ids
        for _ in range(max_new_tokens):
            # crop idx to the last block_size tokens
            idx_cond = idx[:, -self.config.block_size:]
            # get the predictions
            outputs = self(
                input_ids=idx_cond
            )
            logits = outputs.logits
            # 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
            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