"""Load Compactbot/tinystories-50m.

Custom small GPT (not a transformers model). Weights are plain state-dict
tensors in model.safetensors. Tokenizer is a HuggingFace `tokenizers` BPE
file (tokenizer.json).

    from load_model import TinyStoriesGPT, load
    model, tok = load()
    ids = tok.encode("Once upon a time,")
    ...
"""
import torch
import torch.nn as nn
import torch.nn.functional as F
from tokenizers import Tokenizer

D, L, H, FFN, VOCAB, SEQ = 512, 16, 8, 2048, 8192, 512


class RMSNorm(nn.Module):
    def __init__(self, d):
        super().__init__()
        self.w = nn.Parameter(torch.ones(d))
    def forward(self, x):
        return self.w * x * torch.rsqrt(x.float().pow(2).mean(-1, keepdim=True) + 1e-6)


class Block(nn.Module):
    def __init__(self, d, h):
        super().__init__()
        self.ln1 = RMSNorm(d); self.ln2 = RMSNorm(d)
        self.qkv = nn.Linear(d, 3 * d, bias=False)
        self.proj = nn.Linear(d, d, bias=False)
        self.fc1 = nn.Linear(d, FFN, bias=False)
        self.fc2 = nn.Linear(FFN, d, bias=False)
        self.h, self.d = h, d
    def forward(self, x):
        B, T, Dd = x.shape
        h = self.ln1(x)
        qkv = self.qkv(h).view(B, T, 3, self.h, Dd // self.h).transpose(2, 1)
        q, k, v = qkv[:, 0], qkv[:, 1], qkv[:, 2]
        q, k, v = q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2)
        att = F.scaled_dot_product_attention(q, k, v, is_causal=True)
        att = att.transpose(1, 2).reshape(B, T, Dd)
        x = x + self.proj(att)
        x = x + self.fc2(F.gelu(self.fc1(self.ln2(x))))
        return x


class TinyStoriesGPT(nn.Module):
    def __init__(self):
        super().__init__()
        self.tok = nn.Embedding(VOCAB, D)
        self.pos = nn.Embedding(SEQ, D)
        self.blocks = nn.ModuleList([Block(D, H) for _ in range(L)])
        self.ln_f = RMSNorm(D)
        self.lm_head = self.tok  # weight-tied
    def forward(self, idx, targets=None):
        B, T = idx.shape
        x = self.tok(idx) + self.pos(torch.arange(T, device=idx.device))
        for b in self.blocks:
            x = b(x)
        x = self.ln_f(x)
        logits = self.lm_head(x)
        loss = None
        if targets is not None:
            loss = F.cross_entropy(logits.view(-1, VOCAB), targets.view(-1))
        return logits, loss
    def generate(self, idx, max_new, temp=0.8, top_k=40):
        for _ in range(max_new):
            logits, _ = self(idx)
            logits = logits[:, -1] / temp
            if top_k:
                v, _ = torch.topk(logits, top_k)
                logits[logits < v[:, [-1]]] = float("-inf")
            nxt = torch.multinomial(F.softmax(logits, -1), 1)
            idx = torch.cat([idx, nxt], 1)
        return idx


def load(weights="model.safetensors", tokenizer="tokenizer.json", device="cuda"):
    from safetensors.torch import load_file
    m = TinyStoriesGPT().to(device)
    sd = load_file(weights)
    m.load_state_dict(sd)
    m.eval()
    tok = Tokenizer.from_file(tokenizer)
    return m, tok


if __name__ == "__main__":
    m, tok = load()
    ids = tok.encode("Once upon a time,")
    ids = torch.tensor([ids], device="cuda")
    out = m.generate(ids, 100, temp=0.8, top_k=40)
    print(tok.decode(out[0].tolist(), skip_special_tokens=True))