app.py

import torch
import gradio as gr
import tiktoken
from torch import nn
from IPython.display import display, Markdown

# Configuration for the model
GPT_CONFIG_124M = {
    "vocab_size": 50257,  # GPT-2 vocabulary size
    "context_length": 1024,
    "embed_dim": 768,
    "n_layers": 12,
    "n_heads": 12,
    "drop_rate": 0.1,
    "num_experts": 8,  # Number of expert FFNs
    "top_k": 2,  # Number of experts to select per token
    "expert_capacity": 0,
    "qkv_bias": False  # Unlimited capacity by default
}

class GELU(nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, x):
        return 0.5 * x * (1 + torch.tanh(
            torch.sqrt(torch.tensor(2.0 / torch.pi)) * 
            (x + 0.044715 * torch.pow(x, 3))
        ))

class LayerNorm(nn.Module):
    def __init__(self, emb_dim):
        super().__init__()
        self.eps = 1e-5
        self.scale = nn.Parameter(torch.ones(emb_dim))
        self.shift = nn.Parameter(torch.zeros(emb_dim))

    def forward(self, x):
        mean = x.mean(dim=-1, keepdim=True)
        var = x.var(dim=-1, keepdim=True, unbiased=False)
        norm_x = (x - mean) / torch.sqrt(var + self.eps)
        return self.scale * norm_x + self.shift

class FeedForward(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.layers = nn.Sequential(
            nn.Linear(cfg["embed_dim"], 4 * cfg["embed_dim"]),  # Expansion
            GELU(),  # Activation
            nn.Linear(4 * cfg["embed_dim"], cfg["embed_dim"]),  # Contraction
        )

    def forward(self, x):
        return self.layers(x)

class MultiHeadAttention(nn.Module):
    def __init__(self, d_in, d_out, context_length, dropout, num_heads, qkv_bias=False):
        super().__init__()
        assert (d_out % num_heads == 0), "d_out must be divisible by num_heads"
        self.d_out = d_out
        self.num_heads = num_heads
        self.head_dim = d_out // num_heads 

        self.W_query = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.W_key = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.W_value = nn.Linear(d_in, d_out, bias=qkv_bias)
        self.out_proj = nn.Linear(d_out, d_out)  
        self.dropout = nn.Dropout(dropout)
        self.register_buffer(
            "mask", torch.triu(torch.ones(context_length, context_length), diagonal=1)
        )

    def forward(self, x):
        b, num_tokens, d_in = x.shape

        keys = self.W_key(x) 
        queries = self.W_query(x)
        values = self.W_value(x)

        keys = keys.view(b, num_tokens, self.num_heads, self.head_dim) 
        values = values.view(b, num_tokens, self.num_heads, self.head_dim)
        queries = queries.view(b, num_tokens, self.num_heads, self.head_dim)

        keys = keys.transpose(1, 2)
        queries = queries.transpose(1, 2)
        values = values.transpose(1, 2)

        attn_scores = queries @ keys.transpose(2, 3)  

        mask_bool = self.mask.bool()[:num_tokens, :num_tokens]
        attn_scores.masked_fill_(mask_bool, -torch.inf)
        
        attn_weights = torch.softmax(attn_scores / keys.shape[-1]**0.5, dim=-1)
        attn_weights = self.dropout(attn_weights)

        context_vec = (attn_weights @ values).transpose(1, 2) 
        context_vec = context_vec.contiguous().view(b, num_tokens, self.d_out)
        context_vec = self.out_proj(context_vec)

        return context_vec

class TransformerBlock(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.att = MultiHeadAttention(
            d_in=cfg["embed_dim"],
            d_out=cfg["embed_dim"],
            context_length=cfg["context_length"],
            num_heads=cfg["n_heads"], 
            dropout=cfg["drop_rate"],
            qkv_bias=cfg["qkv_bias"]
        )
        self.ff = FeedForward(cfg)
        self.norm1 = LayerNorm(cfg["embed_dim"])
        self.norm2 = LayerNorm(cfg["embed_dim"])
        self.drop_shortcut = nn.Dropout(cfg["drop_rate"])

    def forward(self, x):
        shortcut = x
        x = self.norm1(x)
        x = self.att(x)  
        x = self.drop_shortcut(x)
        x = x + shortcut 

        shortcut = x
        x = self.norm2(x)
        x = self.ff(x)
        x = self.drop_shortcut(x)
        x = x + shortcut 

        return x

class GPTModel(nn.Module):
    def __init__(self, cfg):
        super().__init__()
        self.tok_emb = nn.Embedding(cfg["vocab_size"], cfg["embed_dim"])
        self.pos_emb = nn.Embedding(cfg["context_length"], cfg["embed_dim"])
        self.drop_emb = nn.Dropout(cfg["drop_rate"])
        
        self.trf_blocks = nn.Sequential(
            *[TransformerBlock(cfg) for _ in range(cfg["n_layers"])]
        )
        
        self.final_norm = LayerNorm(cfg["embed_dim"])
        self.out_head = nn.Linear(cfg["embed_dim"], cfg["vocab_size"], bias=False)

    def forward(self, in_idx):
        batch_size, seq_len = in_idx.shape
        tok_embeds = self.tok_emb(in_idx)
        pos_embeds = self.pos_emb(torch.arange(seq_len, device=in_idx.device))
        x = tok_embeds + pos_embeds  
        x = self.drop_emb(x)
        x = self.trf_blocks(x)
        x = self.final_norm(x)
        logits = self.out_head(x)
        return logits

# Define the generate function (inference logic)
def generate(model, idx, max_new_tokens, context_size, temperature=0.0, top_k=None, eos_id=None):
    # Ensure idx has batch dimension
    if idx.dim() == 1:
        idx = idx.unsqueeze(0)
        
    for _ in range(max_new_tokens):
        idx_cond = idx[:, -context_size:]
        with torch.no_grad():
            logits = model(idx_cond)
        logits = logits[:, -1, :]

        if top_k is not None:
            top_logits, _ = torch.topk(logits, top_k)
            min_val = top_logits[:, -1]
            logits = torch.where(logits < min_val, torch.tensor(float("-inf")).to(logits.device), logits)

        if temperature > 0.0:
            logits = logits / temperature
            probs = torch.softmax(logits, dim=-1)
            idx_next = torch.multinomial(probs, num_samples=1)
        else:
            idx_next = torch.argmax(logits, dim=-1, keepdim=True)

        if idx_next == eos_id:
            break

        idx = torch.cat((idx, idx_next), dim=1)

    return idx

# Tokenization helpers
def text_to_token_ids(text, tokenizer):
    encoded = tokenizer.encode('<|endoftext|>', allowed_special={'<|endoftext|>'})[0]
    encoded_tensor = torch.tensor(encoded).unsqueeze(0)
    return encoded_tensor

def token_ids_to_text(token_ids, tokenizer):
    flat = token_ids.squeeze(0)
    return tokenizer.decode(flat.tolist())

# Load model checkpoint
def load_model(checkpoint_path, device, cfg):
    model = GPTModel(cfg)
    model.load_state_dict(torch.load(checkpoint_path, map_location=device))
    model.to(device)
    model.eval()
    return model

# Generate text based on input
def generate_text(input_text, model, tokenizer, device, max_length=100, temperature=0.7, top_k=50, eos_id=None):
    torch.manual_seed(123)  # For reproducibility
    
    input_ids = text_to_token_ids(input_text, tokenizer).to(device)
    
    generated_ids = generate(model, input_ids, max_new_tokens=max_length, context_size=GPT_CONFIG_124M["context_length"], temperature=temperature, top_k=top_k, eos_id=eos_id)
    
    generated_text = token_ids_to_text(generated_ids, tokenizer)
    
    return generated_text

# Gradio Interface
def create_gradio_interface(model, tokenizer, device):
    def gradio_generate(input_text, max_length=100, temperature=0.7, top_k=50):
        eos_id = tokenizer.encode('<|endoftext|>', allowed_special={'<|endoftext|>'})[0]
        return generate_text(input_text, model, tokenizer, device, max_length, temperature, top_k, eos_id)

    interface = gr.Interface(
        fn=gradio_generate,
        inputs=[
            gr.Textbox(label="Enter input text"),
            gr.Slider(minimum=10, maximum=500, step=10, value=100, label="Max Output Length"),
            gr.Slider(minimum=0, maximum=2, step=0.1, value=0.7, label="Temperature"),
            gr.Slider(minimum=1, maximum=100, step=1, value=50, label="Top-k")
        ],
        outputs=gr.Markdown(label="Generated Text"),
        title="Raghu Baba AI yapper",
        description="Enter some input text and generate a yapper response"
    )

    return interface

# Initialize model and tokenizer
tokenizer = tiktoken.get_encoding("gpt2")
device = "cuda" if torch.cuda.is_available() else "cpu"

# Load model checkpoint
checkpoint_path = "dense_gpt_model_checkpoint.pth"  # Path to your model checkpoint
model = load_model(checkpoint_path, device, GPT_CONFIG_124M)

# Create Gradio interface
gradio_interface = create_gradio_interface(model, tokenizer, device)

# Launch the interface
gradio_interface.launch()