model.py

import inspect
import math
from dataclasses import dataclass
from pathlib import Path

import pytorch_lightning as pl
import torch
import torch.nn as nn
from matplotlib import pyplot as plt
from torch.nn import functional as F


@dataclass
class GPTConfig:
    block_size: int = 512
    vocab_size: int = 1024
    n_layer: int = 3
    n_head: int = 3
    n_embd: int = 48
    d_in: int = 5
    dropout: float = 0.1
    bias: bool = True  # True: bias in Linears and LayerNorms, like GPT-2. False: a bit better and faster


class CausalSelfAttention(nn.Module):
    def __init__(self, config: GPTConfig):
        super().__init__()
        assert config.n_embd % config.n_head == 0
        # key, query, value projections for all heads, but in a batch
        self.c_attn = nn.Linear(config.n_embd, 3 * config.n_embd, bias=config.bias)
        # output projection
        self.c_proj = nn.Linear(config.n_embd, config.n_embd, bias=config.bias)
        # regularization
        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()  # batch size, sequence length, embedding dimensionality (n_embd)

        # calculate query, key, values for all heads in batch and move head forward to be the batch dim
        q, k, v = self.c_attn(x).split(self.n_embd, dim=2)
        k = k.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)  # (B, nh, T, hs)
        q = q.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)  # (B, nh, T, hs)
        v = v.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)  # (B, nh, T, hs)

        # causal self-attention; Self-attend: (B, nh, T, hs) x (B, nh, hs, T) -> (B, nh, T, T)
        y = torch.nn.functional.scaled_dot_product_attention(q, k, v, attn_mask=None,
                                                             dropout_p=self.dropout if self.training else 0,
                                                             is_causal=True)
        y = y.transpose(1, 2).contiguous().view(B, T, C)  # re-assemble all head outputs side by side

        # output projection
        y = self.resid_dropout(self.c_proj(y))
        return y


class MLP(nn.Module):
    def __init__(self, config: GPTConfig):
        super().__init__()
        self.c_fc = nn.Linear(config.n_embd, 4 * config.n_embd, bias=config.bias)
        self.tanh = nn.Tanh()
        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.tanh(x)
        x = self.c_proj(x)
        x = self.dropout(x)
        return x


class Block(nn.Module):
    def __init__(self, config: GPTConfig):
        super().__init__()
        self.ln_1 = nn.LayerNorm(config.n_embd, bias=config.bias)
        self.attn = CausalSelfAttention(config)
        self.ln_2 = nn.LayerNorm(config.n_embd, bias=config.bias)
        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: GPTConfig):
        super().__init__()
        assert config.vocab_size is not None
        assert config.block_size is not None
        assert config.n_embd % 2 == 0, "n_embd must be even"
        self.config = config

        self.transformer = nn.ModuleDict(dict(
            wte=nn.Linear(config.d_in - 1, config.n_embd // 2, bias=config.bias),
            wte2=nn.Embedding(config.vocab_size, config.n_embd // 2),
            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, bias=config.bias),
        ))
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        # self.transformer.wte.weight = self.lm_head.weight  # https://paperswithcode.com/method/weight-tying

        # init all weights
        self.apply(self._init_weights)
        # apply special scaled init to the residual projections, per GPT-2 paper
        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, mask=None):
        device = idx.device
        b, t, _ = idx.size()
        assert t <= self.config.block_size, f"Cannot forward sequence of length {t}, block size is only {self.config.block_size}"
        pos = torch.arange(0, t, dtype=torch.long, device=device)  # shape (t)

        # forward the GPT model itself
        num_emb = self.transformer.wte(idx[:, :, :-1])  # numerical "embeddings"
        tok_emb = self.transformer.wte2(idx[:, :, -1].to(dtype=torch.long))  # token embeddings
        emb = torch.cat((num_emb, tok_emb), dim=-1)
        pos_emb = self.transformer.wpe(pos)  # position embeddings of shape (t, n_embd)
        x = self.transformer.drop(emb + pos_emb)
        for block in self.transformer.h:
            x = block(x)
        x = self.transformer.ln_f(x)

        if targets is not None:
            # if we are given some desired targets also calculate the loss
            logits = self.lm_head(x)
            if mask is not None:
                logits = logits[~mask]
                targets = targets[~mask]
            loss = F.cross_entropy(logits.view(-1, logits.size(-1)), targets.view(-1), ignore_index=-1)
        else:
            # inference-time mini-optimization: only forward the lm_head on the very last position
            logits = self.lm_head(x[:, [-1], :])  # note: using list [-1] to preserve the time dim
            loss = None

        return logits, loss

    @torch.no_grad()
    def generate(self, idx, max_new_tokens, exog, temperature=1.0, top_k=None):
        """
        Take a conditioning sequence of indices idx (LongTensor of shape (b,t)) and complete
        the sequence max_new_tokens times, feeding the predictions back into the model each time.
        Most likely you'll want to make sure to be in model.eval() mode of operation for this.
        """
        assert exog.size(2) == self.config.d_in - 1, f"Expected {self.config.d_in - 1} exogenous features, got {exog.size(0)}"
        assert exog.size(1) >= max_new_tokens, f"Expected exogenous features to cover at least {max_new_tokens} time steps, got {exog.size(1)}"
        for i in range(max_new_tokens):
            # if the sequence context is growing too long we must crop it at block_size
            idx_cond = idx if idx.size(1) <= self.config.block_size else idx[:, -self.config.block_size:]
            # forward the model to get the logits for the index in the sequence
            logits, _ = self(idx_cond)
            # pluck the logits at the final step and scale by desired temperature
            logits = logits[:, -1, :] / temperature
            # optionally crop the logits to only the top k options
            if top_k is not None:
                v, _ = torch.topk(logits, min(top_k, logits.size(-1)))
                logits[logits < v[:, [-1]]] = -float('Inf')
            # apply softmax to convert logits to (normalized) probabilities
            probs = F.softmax(logits, dim=-1)
            # sample from the distribution
            idx_next = torch.multinomial(probs, num_samples=1)
            idx_next = torch.cat((exog[:, i, :], idx_next), dim=1).unsqueeze(0)
            # append sampled index to the running sequence and continue
            idx = torch.cat((idx, idx_next), dim=1)

        return idx


class MLControlsSim(pl.LightningModule):
    def __init__(
        self,
        n_layers: int,
        n_head: int,
        n_embd: int,
        lr: float,
        weight_decay: float,
    ):
        super().__init__()
        self.lr = lr
        self.weight_decay = weight_decay
        self.model = GPT(GPTConfig(n_layer=n_layers, n_head=n_head, n_embd=n_embd))
        self.save_hyperparameters()

    def forward(self, *args):
        logits, loss = self.model(*args)
        return logits, loss

    def training_step(self, batch, batch_idx):
        logits, loss = self(*batch)
        self.log('train_loss', loss, on_step=True, prog_bar=True, logger=True)
        return loss

    def validation_step(self, batch, batch_idx):
        logits, loss = self(*batch)
        self.log('val_loss', loss, on_epoch=True, prog_bar=True, logger=True)

        if batch_idx % 5 == 0:
            self.plot_predictions(*batch)
        return loss

    def plot_predictions(self, x, y, mask) -> None:
        dm = self.trainer.datamodule
        idx = x[[0], :dm.CONTEXT_SIZE, :]
        exog = x[[0], dm.CONTEXT_SIZE:, :-1]
        preds = self.model.generate(
            idx=idx,
            exog=exog,
            max_new_tokens=20,
        )
        y_pred = preds[0, :, -1].cpu().numpy().astype(int)
        y_pred = dm.tokenizer.decode(y_pred)
        y_true = y[0, :].cpu().numpy().astype(int)[:len(y_pred)]
        y_true = dm.tokenizer.decode(y_true)

        plt.plot(y_true, label="True")
        plt.plot(y_pred, label="Pred")
        plt.xlabel("Time step")
        plt.ylabel("LatAccel")
        plt.legend()
        save_path = f"{self.logger.log_dir}/val_predictions/{self.current_epoch:03}.png"
        Path(save_path).parent.mkdir(parents=True, exist_ok=True)
        plt.savefig(save_path)
        plt.close()

    def configure_optimizers(self):
        # start with all of the candidate parameters
        param_dict = {pn: p for pn, p in self.named_parameters()}
        # filter out those that do not require grad
        param_dict = {pn: p for pn, p in param_dict.items() if p.requires_grad}
        # create optim groups. Any parameters that is 2D will be weight decayed, otherwise no.
        # i.e. all weight tensors in matmuls + embeddings decay, all biases and layernorms don't.
        decay_params = [p for n, p in param_dict.items() if p.dim() >= 2]
        nodecay_params = [p for n, p in param_dict.items() if p.dim() < 2]
        optim_groups = [
            {'params': decay_params, 'weight_decay': self.weight_decay},
            {'params': nodecay_params, 'weight_decay': 0.0}
        ]
        num_decay_params = sum(p.numel() for p in decay_params)
        num_nodecay_params = sum(p.numel() for p in nodecay_params)
        print(f"num decayed parameter tensors: {len(decay_params)}, with {num_decay_params:,} parameters")
        print(f"num non-decayed parameter tensors: {len(nodecay_params)}, with {num_nodecay_params:,} parameters")
        # Create AdamW optimizer and use the fused version if it is available
        fused_available = 'fused' in inspect.signature(torch.optim.AdamW).parameters
        use_fused = fused_available and self.device.type == 'cuda'
        extra_args = dict(fused=True) if use_fused else dict()
        optimizer = torch.optim.AdamW(optim_groups, lr=self.lr, betas=(0.9, 0.95), **extra_args)
        print(f"using fused AdamW: {use_fused}")

        return optimizer