v1/basic_model_test.py

import torch
import numpy as np
import os
from config import ZUCOConfig
from model import build_model
from transformers import AutoTokenizer


def test_model_basic():
    print("\n" + "="*80)
    print("bai-Mind-8-v1")
    print("="*80 + "\n")

    model_path = r"model/path"

    config = ZUCOConfig()

    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
    print(f"[INFO] Device: {device}")

    print(f"\n[1/4] Creating Model...")
    model = build_model(config)

    total_params = sum(p.numel() for p in model.parameters())
    trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)

    print(f"      Total Params: {total_params:,}")
    print(f"      Trainable Params: {trainable_params:,}")

    print(f"\n[2/4] Model checkpoint loading...")
    print(f"      Path: {model_path}")

    if os.path.exists(model_path):
        checkpoint = torch.load(model_path, map_location=device)
        model.load_state_dict(checkpoint['model_state_dict'])
        print(f"      ✓ Model loaded successfully!")

        if 'epoch' in checkpoint:
            print(f"      Epoch: {checkpoint['epoch']}")
        if 'val_loss' in checkpoint:
            print(f"      Validation Loss: {checkpoint['val_loss']:.4f}")
    else:
        print(f"      ✗ Model path is not found!")
        print(f"      Using randomize model")

    model.to(device)
    model.eval()

    # Tokenizer
    print(f"\n[3/4] Tokenizer loading...")
    tokenizer = AutoTokenizer.from_pretrained(config.language_model_name)
    print(f"      ✓ Tokenizer is ready")

    print(f"\n[4/4] Test...")
    batch_size = 2
    eeg_seq_len = 50

    eeg = torch.randn(batch_size, eeg_seq_len, config.num_eeg_channels).to(device)
    eeg_attention_mask = torch.ones(batch_size, eeg_seq_len).to(device)

    print(f"      Test EEG shape: {eeg.shape}")

    # Inference
    with torch.no_grad():
        generated_ids = model.generate(
            eeg=eeg,
            eeg_attention_mask=eeg_attention_mask,
            max_length=50,
            num_beams=2,
            early_stopping=True
        )

    # Decode
    predictions = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)

    print(f"      ✓ Model is running succesfully!")
    print(f"\n      Sample predictions:")
    for i, pred in enumerate(predictions):
        print(f"        [{i+1}] {pred}")

    print("\n" + "="*80)
    print("TEST RESULT: SUCCESS ✓")
    print("="*80 + "\n")

    return {
        'status': 'success',
        'total_params': total_params,
        'trainable_params': trainable_params,
        'device': str(device),
        'sample_predictions': predictions
    }


if __name__ == "__main__":
    try:
        results = test_model_basic()
        print("[OK] Model test completed!")
    except Exception as e:
        print(f"\n[ERROR]: {e}")
        import traceback
        traceback.print_exc()