Upload 3 files

app.py

@@ -1,8 +1,9 @@
 import torch
 import gradio as gr
 import tiktoken
+import os
 from torch.nn import functional as F
-from train import GPT, GPTConfig
+from model import GPT, GPTConfig
 
 # Initialize model globally to avoid reloading
 model = None
@@ -10,61 +11,103 @@ model = None
 def initialize_model():
     global model
     if model is None:
-        model = GPT(GPTConfig())
-        model.load_state_dict(torch.load('model/model_state_dict.pth', map_location=torch.device('cpu')))
-        model.eval()
+        model_path = 'models/model_state_dict.pth'
+        if not os.path.exists(model_path):
+            raise FileNotFoundError(f"Model file not found at {model_path}")
+        
+        try:
+            model = GPT(GPTConfig())
+            model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu')))
+            model.eval()
+        except Exception as e:
+            raise Exception(f"Error loading model: {str(e)}")
     return model
 
 def generate_shakespeare(prompt, max_length=100, temperature=0.8):
     """Generate Shakespeare-style text from a prompt"""
-    # Initialize model if not already done
-    model = initialize_model()
+    try:
+        # Initialize model if not already done
+        model = initialize_model()
+        
+        # Encode the prompt
+        enc = tiktoken.get_encoding('gpt2')
+        prompt_tokens = enc.encode(prompt)
+        
+        # Safety check for prompt length
+        if len(prompt_tokens) > model.config.block_size:
+            return f"Prompt too long. Please limit to {model.config.block_size} tokens."
+        
+        x = torch.tensor(prompt_tokens).unsqueeze(0)
+        
+        with torch.no_grad():
+            while x.size(1) < max_length:
+                # Get predictions
+                logits, _ = model(x)
+                logits = logits[:, -1, :] / temperature
+                
+                # Sample from the distribution
+                probs = F.softmax(logits, dim=-1)
+                next_token = torch.multinomial(probs, num_samples=1)
+                
+                # Append to the sequence
+                x = torch.cat((x, next_token), dim=1)
+                
+                # Stop if we generate a newline
+                if next_token.item() == enc.encode('\n')[0]:
+                    break
+        
+        # Decode the generated text
+        generated_tokens = x[0].tolist()
+        generated_text = enc.decode(generated_tokens)
+        
+        return generated_text
     
-    # Encode the prompt
-    enc = tiktoken.get_encoding('gpt2')
-    prompt_tokens = enc.encode(prompt)
-    x = torch.tensor(prompt_tokens).unsqueeze(0)
-    
-    with torch.no_grad():
-        while x.size(1) < max_length:
-            # Get predictions
-            logits, _ = model(x)
-            logits = logits[:, -1, :] / temperature
-            
-            # Sample from the distribution
-            probs = F.softmax(logits, dim=-1)
-            next_token = torch.multinomial(probs, num_samples=1)
-            
-            # Append to the sequence
-            x = torch.cat((x, next_token), dim=1)
-            
-            # Stop if we generate a newline
-            if next_token.item() == enc.encode('\n')[0]:
-                break
-    
-    # Decode the generated text
-    generated_tokens = x[0].tolist()
-    generated_text = enc.decode(generated_tokens)
-    
-    return generated_text
+    except Exception as e:
+        return f"Error generating text: {str(e)}"
 
 # Create Gradio interface
 demo = gr.Interface(
     fn=generate_shakespeare,
     inputs=[
-        gr.Textbox(label="Enter your prompt", placeholder="Enter some Shakespeare-style text..."),
-        gr.Slider(minimum=10, maximum=200, value=100, step=1, label="Max Length"),
-        gr.Slider(minimum=0.1, maximum=2.0, value=0.8, step=0.1, label="Temperature")
+        gr.Textbox(
+            label="Enter your prompt",
+            placeholder="Enter some Shakespeare-style text...",
+            lines=2
+        ),
+        gr.Slider(
+            minimum=10,
+            maximum=200,
+            value=100,
+            step=1,
+            label="Max Length",
+            info="Maximum length of generated text"
+        ),
+        gr.Slider(
+            minimum=0.1,
+            maximum=2.0,
+            value=0.8,
+            step=0.1,
+            label="Temperature",
+            info="Higher values make the output more random, lower values make it more focused"
+        )
     ],
-    outputs=gr.Textbox(label="Generated Text"),
+    outputs=gr.Textbox(label="Generated Text", lines=5),
     title="Shakespeare Text Generator",
-    description="Generate Shakespeare-style text based on your prompt using a fine-tuned GPT model.",
+    description="""Generate Shakespeare-style text based on your prompt using a fine-tuned GPT model.
+                   Enter a prompt and adjust the parameters to control the generation.""",
     examples=[
         ["To be, or not to be,", 100, 0.8],
         ["All the world's a stage,", 100, 0.8],
         ["Romeo, Romeo,", 100, 0.8]
-    ]
+    ],
+    cache_examples=True
 )
 
+# Add error handling for the launch
 if __name__ == "__main__":
-    demo.launch() 
+    try:
+        # Test model loading at startup
+        initialize_model()
+        demo.launch()
+    except Exception as e:
+        print(f"Error starting the application: {str(e)}") 

model.py

@@ -0,0 +1,107 @@
+import math
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+from dataclasses import dataclass
+
+@dataclass
+class GPTConfig:
+    block_size: int = 1024
+    vocab_size: int = 50257
+    n_layer: int = 12
+    n_head: int = 12
+    n_embd: int = 768
+
+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)
+        self.c_proj = nn.Linear(config.n_embd, config.n_embd)
+        self.c_proj.NANGPT_SCALE_INIT = 1
+        self.n_head = config.n_head
+        self.n_embd = config.n_embd
+        self.register_buffer("bias", torch.tril(torch.ones(config.block_size, config.block_size)).view(1, 1, config.block_size, config.block_size))
+
+    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)
+        att = (q @ k.transpose(-2, -1)) * (1.0 / math.sqrt(k.size(-1)))
+        att = att.masked_fill(self.bias[:, :, :T, :T] == 0, float('-inf'))
+        att = F.softmax(att, dim=-1)
+        y = att @ v
+        y = y.transpose(1, 2).contiguous().view(B, T, C)
+        y = 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)
+        self.gelu    = nn.GELU(approximate='tanh')
+        self.c_proj  = nn.Linear(4 * config.n_embd, config.n_embd)
+        self.c_proj.NANOGPT_SCALE_INIT = 1
+
+    def forward(self, x):
+        x = self.c_fc(x)
+        x = self.gelu(x)
+        x = self.c_proj(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),
+            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
+        self.apply(self._init_weights)
+
+    def _init_weights(self, module):
+        if isinstance(module, nn.Linear):
+            std = 0.02
+            if hasattr(module, 'NANGPT_SCALE_INIT'):
+                std *= (2 * self.config.n_layer) ** -0.5
+            torch.nn.init.normal_(module.weight, mean=0.0, std=std)
+            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):
+        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=idx.device)
+        pos_emb = self.transformer.wpe(pos)
+        tok_emb = self.transformer.wte(idx)
+        x = tok_emb + pos_emb
+        for block in self.transformer.h:
+            x = block(x)
+        x = self.transformer.ln_f(x)
+        logits = self.lm_head(x)
+        loss = None
+        if targets is not None:
+            loss = F.cross_entropy(logits.view(-1, logits.size(-1)), targets.view(-1))
+        return logits, loss 

train.py

@@ -1,287 +1,287 @@
-# Solving for residual std scaling issue
-import os
-import math
-import time
-import inspect
-from dataclasses import dataclass
-import torch
-import torch.nn as nn
-from torch.nn import functional as F
-
-
-class CausalSelfAttention(nn.Module):
-
-    def __init__(self, config):
-        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)
-        # output projection
-        self.c_proj = nn.Linear(config.n_embd, config.n_embd)
-        self.c_proj.NANGPT_SCALE_INIT = 1
-        # regularization
-        self.n_head = config.n_head
-        self.n_embd = config.n_embd
-        self.register_buffer("bias", torch.tril(torch.ones(config.block_size, config.block_size)).view(1, 1, config.block_size, config.block_size))
-
-    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
-        # nh is "number of heads", hs is "head size", and C (number of channels) = nh * hs
-        # e.g. in GPT-2 (124M), n_head=12, hs=64, so nh*hs=C=768 channels in the Transformer
-        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) # (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)
-
-        att = (q @ k.transpose(-2, -1)) * (1.0 / math.sqrt(k.size(-1)))
-        att = att.masked_fill(self.bias[:, :, :T, :T] == 0, float('-inf'))
-        att = F.softmax(att, dim=-1)
-        y = att @ v # (B, nh, T, T) x (B, nh, T, hs) -> (B, nh, T, hs)
-
-        y = y.transpose(1, 2).contiguous().view(B, T, C) # re-assemble all head outputs side by side
-        # output projection
-        y = 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)
-        self.gelu    = nn.GELU(approximate='tanh')
-        self.c_proj  = nn.Linear(4 * config.n_embd, config.n_embd)
-        self.c_proj.NANOGPT_SCALE_INIT = 1
-
-    def forward(self, x):
-        x = self.c_fc(x)
-        x = self.gelu(x)
-        x = self.c_proj(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
-
-
-@dataclass
-class GPTConfig:
-    block_size: int = 1024 # max sequence length
-    vocab_size: int = 50257 # number of tokens: 50,000 BPE merges + 256 bytes tokens + 1 <|endoftext|> token
-    n_layer: int = 12 # number of layers
-    n_head: int = 12 # number of heads
-    n_embd: int = 768 # embedding dimension
-
-
-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),
-            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)
-
-        # weight sharing
-        self.transformer.wte.weight = self.lm_head.weight
-
-        # weight initialization
-        self.apply(self._init_weights)
-
-    def _init_weights(self, module):
-        if isinstance(module, nn.Linear):
-            std = 0.02
-            if hasattr(module, 'NANGPT_SCALE_INIT'):
-                std *= (2 * self.config.n_layer) ** -0.5
-            torch.nn.init.normal_(module.weight, mean = 0.0, std = std)
-            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):
-        # idx is of shape (B, T)
-        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}"
-        # forward the token and posisition embeddings
-        pos = torch.arange(0, T, dtype=torch.long, device=idx.device) # shape (T)
-        pos_emb = self.transformer.wpe(pos) # position embeddings of shape (T, n_embd)
-        tok_emb = self.transformer.wte(idx) # token embeddings of shape (B, T, n_embd)
-        x = tok_emb + pos_emb
-        # forward the blocks of the transformer
-        for block in self.transformer.h:
-            x = block(x)
-        # forward the final layernorm and the classifier
-        x = self.transformer.ln_f(x)
-        logits = self.lm_head(x) # (B, T, vocab_size)
-        loss = None
-        if targets is not None:
-            loss = F.cross_entropy(logits.view(-1, logits.size(-1)), targets.view(-1))
-        return logits, loss
-
-    @classmethod
-    def from_pretrained(cls, model_type):
-        """Loads pretrained GPT-2 model weights from huggingface"""
-        assert model_type in {'gpt2', 'gpt2-medium', 'gpt2-large', 'gpt2-xl'}
-        from transformers import GPT2LMHeadModel
-        print("loading weights from pretrained gpt: %s" % model_type)
-
-        # n_layer, n_head and n_embd are determined from model_type
-        config_args = {
-            'gpt2':         dict(n_layer=12, n_head=12, n_embd=768),  # 124M params
-            'gpt2-medium':  dict(n_layer=24, n_head=16, n_embd=1024), # 350M params
-            'gpt2-large':   dict(n_layer=36, n_head=20, n_embd=1280), # 774M params
-            'gpt2-xl':      dict(n_layer=48, n_head=25, n_embd=1600), # 1558M params
-        }[model_type]
-        config_args['vocab_size'] = 50257 # always 50257 for GPT model checkpoints
-        config_args['block_size'] = 1024 # always 1024 for GPT model checkpoints
-        # create a from-scratch initialized minGPT model
-        config = GPTConfig(**config_args)
-        model = GPT(config)
-        sd = model.state_dict()
-        sd_keys = sd.keys()
-        sd_keys = [k for k in sd_keys if not k.endswith('.attn.bias')] # discard this mask / buffer, not a param
-
-        # init a huggingface/transformers model
-        model_hf = GPT2LMHeadModel.from_pretrained(model_type)
-        sd_hf = model_hf.state_dict()
-
-        # copy while ensuring all of the parameters are aligned and match in names and shapes
-        sd_keys_hf = sd_hf.keys()
-        sd_keys_hf = [k for k in sd_keys_hf if not k.endswith('.attn.masked_bias')] # ignore these, just a buffer
-        sd_keys_hf = [k for k in sd_keys_hf if not k.endswith('.attn.bias')] # same, just the mask (buffer)
-        transposed = ['attn.c_attn.weight', 'attn.c_proj.weight', 'mlp.c_fc.weight', 'mlp.c_proj.weight']
-        # basically the openai checkpoints use a "Conv1D" module, but we only want to use a vanilla Linear
-        # this means that we have to transpose these weights when we import them
-        assert len(sd_keys_hf) == len(sd_keys), f"mismatched keys: {len(sd_keys_hf)} != {len(sd_keys)}"
-        for k in sd_keys_hf:
-            if any(k.endswith(w) for w in transposed):
-                # special treatment for the Conv1D weights we need to transpose
-                assert sd_hf[k].shape[::-1] == sd[k].shape
-                with torch.no_grad():
-                    sd[k].copy_(sd_hf[k].t())
-            else:
-                # vanilla copy over the other parameters
-                assert sd_hf[k].shape == sd[k].shape
-                with torch.no_grad():
-                    sd[k].copy_(sd_hf[k])
-
-        return model
-
-# model = GPT.from_pretrained('gpt2')
-
-device = 'cpu'
-if torch.cuda.is_available():
-    device = 'cuda'
-elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available():
-    device = "mps"
-print(f"using device: {device}")
-
-# SEED
-torch.manual_seed(1337)
-if torch.cuda.is_available():
-    torch.cuda.manual_seed(1337)
-
-# STOP
-num_return_sequences = 5
-max_length = 30
-
-
-
-import tiktoken
-
-class DataLoaderLite:
-    def __init__(self, B, T):
-        self.B = B
-        self.T = T
-
-        # at init load tokens from disk and store them in memory
-        with open('data/input.txt', 'r') as f:
-            text = f.read()
-        enc = tiktoken.get_encoding('gpt2') 
-        tokens = enc.encode(text)
-        self.tokens = torch.tensor(tokens)
-        print(f'loaded {len(self.tokens)} tokens')
-        print(f'1 epoch = {len(self.tokens) // (B * T)} batches')
-
-        # state
-        self.current_position = 0
-    
-    def next_batch(self):
-        B, T = self.B, self.T
-        buf = self.tokens[self.current_position: self.current_position + B * T + 1]
-        x = (buf[:-1]).view(B, T) # inputs
-        y = (buf[1:]).view(B, T) # targets
-        # advance the position in the tensor
-        self.current_position += B*T
-        # if loading the next batch would be out of bounds, reset
-        if self.current_position + (B * T + 1) > len(self.tokens):
-            self.current_position = 0
-        return x, y
-
-
-model = GPT(GPTConfig())
-model.to(device)
-
-train_loader = DataLoaderLite(B = 4, T = 32)
-
-# NEW CODE
-optimizer = torch.optim.AdamW(model.parameters(), lr = 3e-4)
-for i in range(50):
-    x, y = train_loader.next_batch()
-    x, y = x.to(device), y.to(device)
-    optimizer.zero_grad()
-    logits, loss = model(x, y)
-    loss.backward()
-    optimizer.step()
-    print(f'step{i}, loss: {loss.item()}')
-
-
-print(loss)
-import sys; sys.exit(0)
-
-torch.manual_seed(42)
-torch.cuda.manual_seed(42)
-while x.size(1) < max_length:
-    # forward the model to get the logits
-    with torch.no_grad():
-        logits = model(x)[0] # (B, T, vocab_size)
-        # take the logits at the last position
-        logits = logits[:, -1, :] # (B, vocab_size)
-        # get the probabilities
-        probs = F.softmax(logits, dim=-1)
-        # do top-k sampling of 50 (huggingface pipeline default)
-        # topk_probs here becomes (5, 50), topk_indices is (5, 50)
-        topk_probs, topk_indices = torch.topk(probs, 50, dim=-1)
-        # select a token from the top-k probabilities
-        # note: multinomial does not demand the input to sum to 1
-        ix = torch.multinomial(topk_probs, 1) # (B, 1)
-        # gather the corresponding indices
-        xcol = torch.gather(topk_indices, -1, ix) # (B, 1)
-        # append to the sequence
-        x = torch.cat((x, xcol), dim=1)
-
-# print the generated text
-for i in range(num_return_sequences):
-    tokens = x[i, :max_length].tolist()
-    decoded = enc.decode(tokens)
+# Solving for residual std scaling issue
+import os
+import math
+import time
+import inspect
+from dataclasses import dataclass
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+
+
+class CausalSelfAttention(nn.Module):
+
+    def __init__(self, config):
+        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)
+        # output projection
+        self.c_proj = nn.Linear(config.n_embd, config.n_embd)
+        self.c_proj.NANGPT_SCALE_INIT = 1
+        # regularization
+        self.n_head = config.n_head
+        self.n_embd = config.n_embd
+        self.register_buffer("bias", torch.tril(torch.ones(config.block_size, config.block_size)).view(1, 1, config.block_size, config.block_size))
+
+    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
+        # nh is "number of heads", hs is "head size", and C (number of channels) = nh * hs
+        # e.g. in GPT-2 (124M), n_head=12, hs=64, so nh*hs=C=768 channels in the Transformer
+        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) # (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)
+
+        att = (q @ k.transpose(-2, -1)) * (1.0 / math.sqrt(k.size(-1)))
+        att = att.masked_fill(self.bias[:, :, :T, :T] == 0, float('-inf'))
+        att = F.softmax(att, dim=-1)
+        y = att @ v # (B, nh, T, T) x (B, nh, T, hs) -> (B, nh, T, hs)
+
+        y = y.transpose(1, 2).contiguous().view(B, T, C) # re-assemble all head outputs side by side
+        # output projection
+        y = 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)
+        self.gelu    = nn.GELU(approximate='tanh')
+        self.c_proj  = nn.Linear(4 * config.n_embd, config.n_embd)
+        self.c_proj.NANOGPT_SCALE_INIT = 1
+
+    def forward(self, x):
+        x = self.c_fc(x)
+        x = self.gelu(x)
+        x = self.c_proj(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
+
+
+@dataclass
+class GPTConfig:
+    block_size: int = 1024 # max sequence length
+    vocab_size: int = 50257 # number of tokens: 50,000 BPE merges + 256 bytes tokens + 1 <|endoftext|> token
+    n_layer: int = 12 # number of layers
+    n_head: int = 12 # number of heads
+    n_embd: int = 768 # embedding dimension
+
+
+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),
+            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)
+
+        # weight sharing
+        self.transformer.wte.weight = self.lm_head.weight
+
+        # weight initialization
+        self.apply(self._init_weights)
+
+    def _init_weights(self, module):
+        if isinstance(module, nn.Linear):
+            std = 0.02
+            if hasattr(module, 'NANGPT_SCALE_INIT'):
+                std *= (2 * self.config.n_layer) ** -0.5
+            torch.nn.init.normal_(module.weight, mean = 0.0, std = std)
+            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):
+        # idx is of shape (B, T)
+        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}"
+        # forward the token and posisition embeddings
+        pos = torch.arange(0, T, dtype=torch.long, device=idx.device) # shape (T)
+        pos_emb = self.transformer.wpe(pos) # position embeddings of shape (T, n_embd)
+        tok_emb = self.transformer.wte(idx) # token embeddings of shape (B, T, n_embd)
+        x = tok_emb + pos_emb
+        # forward the blocks of the transformer
+        for block in self.transformer.h:
+            x = block(x)
+        # forward the final layernorm and the classifier
+        x = self.transformer.ln_f(x)
+        logits = self.lm_head(x) # (B, T, vocab_size)
+        loss = None
+        if targets is not None:
+            loss = F.cross_entropy(logits.view(-1, logits.size(-1)), targets.view(-1))
+        return logits, loss
+
+    @classmethod
+    def from_pretrained(cls, model_type):
+        """Loads pretrained GPT-2 model weights from huggingface"""
+        assert model_type in {'gpt2', 'gpt2-medium', 'gpt2-large', 'gpt2-xl'}
+        from transformers import GPT2LMHeadModel
+        print("loading weights from pretrained gpt: %s" % model_type)
+
+        # n_layer, n_head and n_embd are determined from model_type
+        config_args = {
+            'gpt2':         dict(n_layer=12, n_head=12, n_embd=768),  # 124M params
+            'gpt2-medium':  dict(n_layer=24, n_head=16, n_embd=1024), # 350M params
+            'gpt2-large':   dict(n_layer=36, n_head=20, n_embd=1280), # 774M params
+            'gpt2-xl':      dict(n_layer=48, n_head=25, n_embd=1600), # 1558M params
+        }[model_type]
+        config_args['vocab_size'] = 50257 # always 50257 for GPT model checkpoints
+        config_args['block_size'] = 1024 # always 1024 for GPT model checkpoints
+        # create a from-scratch initialized minGPT model
+        config = GPTConfig(**config_args)
+        model = GPT(config)
+        sd = model.state_dict()
+        sd_keys = sd.keys()
+        sd_keys = [k for k in sd_keys if not k.endswith('.attn.bias')] # discard this mask / buffer, not a param
+
+        # init a huggingface/transformers model
+        model_hf = GPT2LMHeadModel.from_pretrained(model_type)
+        sd_hf = model_hf.state_dict()
+
+        # copy while ensuring all of the parameters are aligned and match in names and shapes
+        sd_keys_hf = sd_hf.keys()
+        sd_keys_hf = [k for k in sd_keys_hf if not k.endswith('.attn.masked_bias')] # ignore these, just a buffer
+        sd_keys_hf = [k for k in sd_keys_hf if not k.endswith('.attn.bias')] # same, just the mask (buffer)
+        transposed = ['attn.c_attn.weight', 'attn.c_proj.weight', 'mlp.c_fc.weight', 'mlp.c_proj.weight']
+        # basically the openai checkpoints use a "Conv1D" module, but we only want to use a vanilla Linear
+        # this means that we have to transpose these weights when we import them
+        assert len(sd_keys_hf) == len(sd_keys), f"mismatched keys: {len(sd_keys_hf)} != {len(sd_keys)}"
+        for k in sd_keys_hf:
+            if any(k.endswith(w) for w in transposed):
+                # special treatment for the Conv1D weights we need to transpose
+                assert sd_hf[k].shape[::-1] == sd[k].shape
+                with torch.no_grad():
+                    sd[k].copy_(sd_hf[k].t())
+            else:
+                # vanilla copy over the other parameters
+                assert sd_hf[k].shape == sd[k].shape
+                with torch.no_grad():
+                    sd[k].copy_(sd_hf[k])
+
+        return model
+
+# model = GPT.from_pretrained('gpt2')
+
+device = 'cpu'
+if torch.cuda.is_available():
+    device = 'cuda'
+elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available():
+    device = "mps"
+print(f"using device: {device}")
+
+# SEED
+torch.manual_seed(1337)
+if torch.cuda.is_available():
+    torch.cuda.manual_seed(1337)
+
+# STOP
+num_return_sequences = 5
+max_length = 30
+
+
+
+import tiktoken
+
+class DataLoaderLite:
+    def __init__(self, B, T):
+        self.B = B
+        self.T = T
+
+        # at init load tokens from disk and store them in memory
+        with open('data/input.txt', 'r') as f:
+            text = f.read()
+        enc = tiktoken.get_encoding('gpt2') 
+        tokens = enc.encode(text)
+        self.tokens = torch.tensor(tokens)
+        print(f'loaded {len(self.tokens)} tokens')
+        print(f'1 epoch = {len(self.tokens) // (B * T)} batches')
+
+        # state
+        self.current_position = 0
+    
+    def next_batch(self):
+        B, T = self.B, self.T
+        buf = self.tokens[self.current_position: self.current_position + B * T + 1]
+        x = (buf[:-1]).view(B, T) # inputs
+        y = (buf[1:]).view(B, T) # targets
+        # advance the position in the tensor
+        self.current_position += B*T
+        # if loading the next batch would be out of bounds, reset
+        if self.current_position + (B * T + 1) > len(self.tokens):
+            self.current_position = 0
+        return x, y
+
+
+model = GPT(GPTConfig())
+model.to(device)
+
+train_loader = DataLoaderLite(B = 4, T = 32)
+
+# NEW CODE
+optimizer = torch.optim.AdamW(model.parameters(), lr = 3e-4)
+for i in range(50):
+    x, y = train_loader.next_batch()
+    x, y = x.to(device), y.to(device)
+    optimizer.zero_grad()
+    logits, loss = model(x, y)
+    loss.backward()
+    optimizer.step()
+    print(f'step{i}, loss: {loss.item()}')
+
+
+print(loss)
+import sys; sys.exit(0)
+
+torch.manual_seed(42)
+torch.cuda.manual_seed(42)
+while x.size(1) < max_length:
+    # forward the model to get the logits
+    with torch.no_grad():
+        logits = model(x)[0] # (B, T, vocab_size)
+        # take the logits at the last position
+        logits = logits[:, -1, :] # (B, vocab_size)
+        # get the probabilities
+        probs = F.softmax(logits, dim=-1)
+        # do top-k sampling of 50 (huggingface pipeline default)
+        # topk_probs here becomes (5, 50), topk_indices is (5, 50)
+        topk_probs, topk_indices = torch.topk(probs, 50, dim=-1)
+        # select a token from the top-k probabilities
+        # note: multinomial does not demand the input to sum to 1
+        ix = torch.multinomial(topk_probs, 1) # (B, 1)
+        # gather the corresponding indices
+        xcol = torch.gather(topk_indices, -1, ix) # (B, 1)
+        # append to the sequence
+        x = torch.cat((x, xcol), dim=1)
+
+# print the generated text
+for i in range(num_return_sequences):
+    tokens = x[i, :max_length].tolist()
+    decoded = enc.decode(tokens)
     print(">", decoded)