generate.py

"""
Text Generation - Small Language Model
=======================================
Author: Jekardah AI Lab

Usage:
    python generate.py
    python generate.py --prompt "indonesia adalah"
    python generate.py --interactive
"""

import os
import sys
import argparse
import torch

sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "bpe-tokenizer-id"))
from model import SmallLM
from bpe_tokenizer import BPETokenizer


def load_model(model_dir="./output_slm"):
    model = SmallLM.from_pretrained(model_dir)
    tokenizer = BPETokenizer.from_pretrained(model_dir)
    return model, tokenizer


def generate(model, tokenizer, prompt, max_tokens=50, temperature=0.8, top_k=40):
    ids = tokenizer.encode(prompt.lower())
    inp = torch.tensor([ids], dtype=torch.long)
    out = model.generate(inp, max_new_tokens=max_tokens,
                         temperature=temperature, top_k=top_k)
    return tokenizer.decode(out[0].tolist())


def demo(model, tokenizer):
    print("\n" + "=" * 60)
    print("🇮🇩 SLM Bahasa Indonesia - Demo")
    print("=" * 60)

    prompts = [
        "indonesia adalah",
        "pendidikan",
        "jakarta",
        "makan nasi",
        "teknologi",
        "kebudayaan indonesia",
        "demokrasi",
        "ekonomi",
        "saya suka",
        "hutan",
    ]

    for p in prompts:
        text = generate(model, tokenizer, p, max_tokens=30)
        print(f"\n📝 \"{p}\"")
        print(f"🤖 {text[:80]}")

    print("\n" + "=" * 60)


def interactive(model, tokenizer):
    print("\n" + "=" * 60)
    print("🇮🇩 SLM Bahasa Indonesia - Interactive")
    print("=" * 60)
    print("Ketik awalan kalimat, model melanjutkan.")
    print("Ketik 'quit' untuk keluar.\n")

    while True:
        try:
            prompt = input("📝 Input: ").strip()
        except (EOFError, KeyboardInterrupt):
            break
        if prompt.lower() in ('quit', 'exit', 'q'):
            break
        if not prompt:
            continue

        text = generate(model, tokenizer, prompt, max_tokens=50)
        print(f"🤖 {text}\n")

    print("👋 Bye!")


def main():
    parser = argparse.ArgumentParser(description="SLM Generator")
    parser.add_argument("--model-dir", default="./output_slm")
    parser.add_argument("--prompt", type=str)
    parser.add_argument("--interactive", action="store_true")
    parser.add_argument("--max-tokens", type=int, default=50)
    parser.add_argument("--temperature", type=float, default=0.8)
    args = parser.parse_args()

    model, tokenizer = load_model(args.model_dir)

    if args.interactive:
        interactive(model, tokenizer)
    elif args.prompt:
        print(generate(model, tokenizer, args.prompt,
                       max_tokens=args.max_tokens, temperature=args.temperature))
    else:
        demo(model, tokenizer)


if __name__ == "__main__":
    main()