Update gpt.py

gpt.py

@@ -1,138 +1,120 @@
 import torch
-import torch.nn as nn
-from torch.nn import functional as F
-import config as cfg
+from torch import nn
+import torch.nn.functional as F
 
-class Head(nn.Module):
-    """ one head of self-attention """
 
-    def __init__(self, head_size):
+class Head(nn.Module):
+    def __init__(self, n_embeds, head_size, block_size, dropout) -> None:
         super().__init__()
-        self.key = nn.Linear(cfg.n_embd, head_size, bias=False)
-        self.query = nn.Linear(cfg.n_embd, head_size, bias=False)
-        self.value = nn.Linear(cfg.n_embd, head_size, bias=False)
-        self.register_buffer('tril', torch.tril(torch.ones(cfg.block_size, cfg.block_size)))
-
-        self.dropout = nn.Dropout(cfg.dropout)
+        self.key = nn.Linear(n_embeds, head_size, bias=False)
+        self.query = nn.Linear(n_embeds, head_size, bias=False)
+        self.value = nn.Linear(n_embeds, head_size, bias=False)
+        self.dropout = nn.Dropout(dropout)
+        self.register_buffer("tril", torch.tril(torch.ones(block_size, block_size)))
 
     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)
+        B, T, C = x.shape
+        k = self.key(x)
+        q = self.query(x)
+        wei = q @ k.transpose(-2, -1) * (C**-0.5)  # (B,T,16) @ (B,16,T) --> (B,T,T)
+        wei = wei.masked_fill(self.tril[:T, :T] == 0, float("-inf"))
+        wei = F.softmax(wei, dim=-1)
         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)
+        v = self.value(x)
+        out = wei @ v
         return out
 
-class MultiHeadAttention(nn.Module):
-    """ multiple heads of self-attention in parallel """
 
-    def __init__(self, num_heads, head_size):
+class MultiHeadAttention(nn.Module):
+    def __init__(self, n_heads, n_embeds, head_size, block_size, dropout):
         super().__init__()
-        self.heads = nn.ModuleList([Head(head_size) for _ in range(num_heads)])
-        self.proj = nn.Linear(head_size * num_heads, cfg.n_embd)
-        self.dropout = nn.Dropout(cfg.dropout)
+        self.heads = nn.ModuleList(
+            [Head(n_embeds, head_size, block_size, dropout) for _ in range(n_heads)]
+        )
+        self.proj = nn.Linear(n_embeds, n_embeds)
+        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
+        x = torch.cat([h(x) for h in self.heads], dim=-1)
+        x = self.proj(x)
+        x = self.dropout(x)
+        return x
 
-class FeedFoward(nn.Module):
-    """ a simple linear layer followed by a non-linearity """
 
-    def __init__(self, n_embd):
+class FeedForward(nn.Module):
+    def __init__(self, n_embeds, dropout):
         super().__init__()
         self.net = nn.Sequential(
-            nn.Linear(n_embd, 4 * n_embd),
+            nn.Linear(n_embeds, 4 * n_embeds),
             nn.ReLU(),
-            nn.Linear(4 * n_embd, n_embd),
-            nn.Dropout(cfg.dropout),
+            nn.Linear(4 * n_embeds, n_embeds),
+            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
+class Decoder(nn.Module):
+    def __init__(self, n_embeds, n_heads, block_size, dropout):
         super().__init__()
-        head_size = n_embd // n_head
-        self.sa = MultiHeadAttention(n_head, head_size)
-        self.ffwd = FeedFoward(n_embd)
-        self.ln1 = nn.LayerNorm(n_embd)
-        self.ln2 = nn.LayerNorm(n_embd)
+        head_size = n_embeds // n_heads
+        self.sa_heads = MultiHeadAttention(
+            n_heads, n_embeds, head_size, block_size, dropout
+        )
+        self.ffwd = FeedForward(n_embeds, dropout)
+        self.ln1 = nn.LayerNorm(n_embeds)
+        self.ln2 = nn.LayerNorm(n_embeds)
 
     def forward(self, x):
-        x = x + self.sa(self.ln1(x))
+        x = x + self.sa_heads(self.ln1(x))
         x = x + self.ffwd(self.ln2(x))
         return x
 
-class GPTLanguageModel(nn.Module):
 
-    def __init__(self, vocab_size):
+class GPTModel(nn.Module):
+    def __init__(
+        self, vocab_size, n_embeds, block_size, n_heads, n_layers, dropout, device
+    ):
         super().__init__()
-        # each token directly reads off the logits for the next token from a lookup table
-        self.token_embedding_table = nn.Embedding(vocab_size, cfg.n_embd)
-        self.position_embedding_table = nn.Embedding(cfg.block_size, cfg.n_embd)
-        self.blocks = nn.Sequential(*[Block(cfg.n_embd, n_head=cfg.n_head) for _ in range(cfg.n_layer)])
-        self.ln_f = nn.LayerNorm(cfg.n_embd) # final layer norm
-        self.lm_head = nn.Linear(cfg.n_embd, vocab_size)
-
-        # better init, not covered in the original GPT video, but important, will cover in followup video
-        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)
+        self.device = device
+        self.block_size = block_size
+        self.token_embedding_table = nn.Embedding(vocab_size, n_embeds)
+        self.position_embedding_table = nn.Embedding(block_size, n_embeds)
+        self.blocks = nn.Sequential(
+            *[Decoder(n_embeds, n_heads, block_size, dropout) for _ in range(n_layers)]
+        )
+        self.lnf = nn.LayerNorm(n_embeds)
+        self.lm_head = nn.Linear(n_embeds, vocab_size)
 
     def forward(self, idx, targets=None):
         B, T = idx.shape
 
-        # idx and targets are both (B,T) tensor of integers
-        tok_emb = self.token_embedding_table(idx) # (B,T,C)
-        pos_emb = self.position_embedding_table(torch.arange(T, device=cfg.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:
+        tok_embeds = self.token_embedding_table(idx)  # BxTxNemb
+        pos_embeds = self.position_embedding_table(
+            torch.arange(T, device=self.device)
+        )  # TXNemb
+
+        x = tok_embeds + pos_embeds  # BxTxNemb
+        x = self.blocks(x)
+        x = self.lnf(x)
+        logits = self.lm_head(x)  # BxTxVocabSize
+
+        loss = None
+        if targets is not None:
             B, T, C = logits.shape
-            logits = logits.view(B*T, C)
-            targets = targets.view(B*T)
+            logits = logits.view(B * T, C)
+            targets = targets.view(B * T)
             loss = F.cross_entropy(logits, targets)
-
         return logits, loss
 
     def generate(self, idx, max_new_tokens):
-        # idx is (B, T) array of indices in the current context
         for _ in range(max_new_tokens):
-            # crop idx to the last block_size tokens
-            idx_cond = idx[:, -cfg.block_size:]
-            # get the predictions
-            logits, loss = self(idx_cond)
-            # 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)
+            idx_cond = idx[:, -self.block_size :]
+            logits, loss = self(idx_cond)  # BxTxC
+            logits = logits[:, -1, :]  # BxC
+            probs = F.softmax(logits, dim=-1)  # BxC
+            idx_next = torch.multinomial(probs, num_samples=1)  # Bx1
+            idx = torch.cat((idx, idx_next), dim=1)  # BxT+1
+
         return idx