app.py

import gradio as gr

import gdown
import torch
import torch.nn as nn
from torch.nn import functional as F

device = "cuda" if torch.cuda.is_available() else "cpu"
n_embd = 384
n_head = 4
n_layer = 4
block_size = 128
dropout = 0.2


gdown.download('https://drive.usercontent.google.com/download?id=14k2xUrvJ32trhLCzV2_O7klreBBA3dUu&authuser=0&confirm=t', 'model.pth', quiet=False)
gdown.download('https://drive.usercontent.google.com/download?id=1-JSvTzTxyI5zJwO39o0wuxJpvY-NqzGE&export=download&authuser=0&confirm=t&uuid=9eff48e6-67f8-4728-aa7f-552c497fb02c&at=AN_67v0xah9SgNOs5FDNKIuxVWL9%3A1727637766874', 'data.txt.gz', quiet=False)


import gzip
with gzip.open('data.txt.gz', 'rt', encoding='utf-8') as f:
    dataset = f.read()
#    chars = sorted(list(set(dataset)))
chars = ['\t', '\n', ' ', '!', '"', '#', '$', '%', '&', "'", '(', ')', '*', '+', ',', '-', '.', '/', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', ':', ';', '<', '=', '>', '?', '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', '{', '|', '}', '~', '\x81', '\x8d', '\x8f', '\x90', '\x92', '\x93', '\x94', '\x9d', '\xa0', '¡', '¢', '£', '¤', '¥', '¦', '§', '¨', '©', 'ª', '«', '¬', '\xad', '®', '¯', '°', '±', '²', '³', '´', 'µ', '¶', '·', '¸', '¹', 'º', '»', '¼', '½', '¾', '¿', 'Â', 'Ã', 'Æ', 'Ç', 'É', 'Ê', 'Ë', 'Ð', 'Ò', '×', 'Ø', 'Ù', 'à', 'á', 'â', 'ã', 'ä', 'å', 'é', 'í', 'ï', 'ð', 'ñ', 'ó', 'ö', 'ā', 'Œ', 'œ', 'Š', 'š', 'Ÿ', 'Ž', 'ž', 'ƒ', 'ˆ', '˜', 'і', '\u2005', '\u2009', '\u200a', '\u200b', '\u200e', '–', '—', '―', '‘', '’', '‚', '“', '”', '„', '†', '‡', '•', '…', '\u2028', '\u2029', '\u202a', '‰', '′', '‹', '›', '€', '™', '−', '─', '」', 'fi', '\ufeff', '�', '𝑐', '🌴', '🌹', '🍌', '🙂']
vocab_size = 212# len(chars)

string_to_int = { ch:i for i,ch in enumerate(chars) }
int_to_string = { i:ch for i,ch in enumerate(chars) }
encode = lambda s: [string_to_int[c] for c in s]
decode = lambda l: ''.join([int_to_string[i] for i in l])

data = torch.tensor(encode(dataset), dtype=torch.long)

class Head(nn.Module):

    def __init__(self, head_size):
        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)
        q = self.query(x)
        w = q @ k.transpose(-2, -1) * C**-0.5
        w = w.masked_fill(self.tril[:T, :T] == 0, float("-inf"))
        w = F.softmax(w, dim=-1)
        w = self.dropout(w)
        v = self.value(x)
        out = w @ v
        return out


class MultiHeadAttention(nn.Module):

    def __init__(self, num_heads, head_size):
        super().__init__()
        self.heads = nn.ModuleList([Head(head_size) 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):

    def __init__(self, n_embd):
        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):

    def __init__(self, n_embd, n_head):
        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)

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


class GPTLanguageModel(nn.Module):
    def __init__(self, vocab_size):
        super().__init__()
        self.token_embedding_table = nn.Embedding(vocab_size, n_embd)
        self.position_embedding_table = nn.Embedding(block_size, n_embd)
        self.blocks = nn.Sequential(
            *[Block(n_embd, n_head=n_head) for _ in range(n_layer)]
        )
        self.ln_f = nn.LayerNorm(n_embd)
        self.lm_head = nn.Linear(n_embd, vocab_size)

    def forward(self, idx, targets=None):
        B, T = idx.shape

        tok_emb = self.token_embedding_table(idx)
        pos_emb = self.position_embedding_table(torch.arange(T, device=device))
        x = tok_emb + pos_emb
        x = self.blocks(x)
        x = self.ln_f(x)
        logits = self.lm_head(x)

        if targets is None:
            loss = None
        else:
            B, T, C = logits.shape
            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):
        for _ in range(max_new_tokens):
            idx_cond = idx[:, -block_size:]
            logits, loss = self(idx_cond)
            logits = logits[:, -1, :]
            probs = F.softmax(logits, dim=-1)
            idx_next = torch.multinomial(probs, num_samples=1)
            idx = torch.cat((idx, idx_next), dim=1)
        return idx


@torch.no_grad()
def estimate_loss():
    out = {}
    model.eval()
    for split in ["training", "validation"]:
        losses = torch.zeros(eval_iters)
        for k in range(eval_iters):
            X, Y = get_batch(split)
            logits, loss = model(X, Y)
            losses[k] = loss.item()
        out[split] = losses.mean()
    model.train()
    return out


model = GPTLanguageModel(vocab_size)
m = model.to(device)
# print(sum(p.numel() for p in m.parameters()) / 1e3, "K parameters")

# load the model.pth
model.load_state_dict(torch.load("model.pth", map_location=torch.device('cpu')), strict=False)

model.eval()

def respond(
    message,
    max_tokens=512,
):
    context = torch.tensor(encode(message), dtype=torch.long, device=device).unsqueeze(
        0
    )
    response = decode(model.generate(context, max_new_tokens=max_tokens)[0].tolist())
    return response


iface = gr.Interface(
    fn=respond,
    inputs=[
        gr.Textbox(lines=5, label="Message", value="Hi Harry Potter"),
        gr.Slider(minimum=100, maximum=2048, value=256, label="Max Tokens"),
    ],
    outputs="text",
    title="PotterLLM",
    description="A language model trained on Harry Potter Series.",
    theme="huggingface",
)

if __name__ == "__main__":
    iface.launch()