inference.py

from huggingface_hub import hf_hub_download
import torch
from torch import nn
from torch.nn import functional as F
from torch.utils.data import Dataset, DataLoader, random_split
import urllib.request
import os
from transformers import AutoTokenizer, logging
import pandas as pd
from tqdm import tqdm
from safetensors.torch import load_file


class TransformerBlock(nn.Module):
    def __init__(self, emb_dim, num_heads, context_length, dropout=0.1):
        super().__init__()
        self.ln1 = nn.LayerNorm(emb_dim)
        self.ln2 = nn.LayerNorm(emb_dim)
        self.attn = nn.MultiheadAttention(
            emb_dim, num_heads, dropout=dropout, batch_first=True
        )
        self.mlp = nn.Sequential(
            nn.Linear(emb_dim, 4 * emb_dim),
            nn.GELU(),
            nn.Linear(4 * emb_dim, emb_dim),
            nn.Dropout(dropout),
        )

    def forward(self, x):
        attn_out, _ = self.attn(
            self.ln1(x), self.ln1(x), self.ln1(x), need_weights=False
        )
        x = x + attn_out
        x = x + self.mlp(self.ln2(x))
        return x


class MiniTransformer(nn.Module):
    def __init__(
        self,
        vocab_size,
        emb_dim,
        context_length,
        num_heads,
        num_layers,
        dropout=0.1,
    ):
        super().__init__()
        self.emb = nn.Embedding(vocab_size, emb_dim)
        self.pos_emb = nn.Embedding(context_length, emb_dim)
        self.blocks = nn.Sequential(
            *[
                TransformerBlock(emb_dim, num_heads, context_length, dropout)
                for _ in range(num_layers)
            ]
        )
        self.ln_f = nn.LayerNorm(emb_dim)
        self.head = nn.Linear(emb_dim, vocab_size, bias=False)
        self.context_length = context_length

    def forward(self, x):
        B, T = x.shape
        pos = torch.arange(T, device=x.device)
        x = self.emb(x) + self.pos_emb(pos)
        x = self.blocks(x)
        x = self.ln_f(x)
        logits = self.head(x)
        return logits

    @torch.no_grad()
    def generate(self, x, max_new_tokens=20, temperature=1.0, top_k=None):

        for _ in range(max_new_tokens):
            # truncate context if needed
            x_cond = x[:, -self.context_length :]

            # get predictions
            logits = self(x_cond)  # (B, T_cond, vocab_size)
            logits = logits[:, -1, :] / temperature  # only last position

            # optionally restrict to top-k

            probs = F.softmax(logits, dim=-1)

            # sample from the distribution
            next_token = torch.multinomial(probs, num_samples=1)  # (B, 1)
            # next_token = torch.argmax(probs, dim = 1).unsqueeze(-1)
            # append to sequence
            x = torch.cat([x, next_token], dim=1)

        return x


CONTEXT_LENGTH = 128
EMBEDDING_DIMENSION = 512
HEAD_NUMBER = 4
N_LAYER = 4
tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
device = torch.device("cuda" if torch.cuda.is_available() else "mps")

# Download the model file
model_path = hf_hub_download(
    repo_id="pierjoe/MiniTransformer",
    filename="checkpoints/mini_transformer_v3/model_40.safetensors",
)

# Load with your custom class
model = MiniTransformer(
    vocab_size=tokenizer.vocab_size,
    emb_dim=EMBEDDING_DIMENSION,
    context_length=CONTEXT_LENGTH,
    num_heads=HEAD_NUMBER,
    num_layers=N_LAYER,
).to(device)
state_dict = load_file(model_path)
state_dict = {k.replace("_orig_mod.", ""): v for k, v in state_dict.items()}

model.load_state_dict(state_dict)

model.eval()
max_tokens = 100
prompt = "You are a helpful assistant. Provide clear, concise, and accurate responses to the user "
input_ids = tokenizer.encode(prompt, return_tensors="pt").to(device)
output_ids = model.generate(
    input_ids, max_new_tokens=max_tokens, temperature=5, top_k=10
)
generated_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
generated_text