test.py

import CNN
import torch
from FEN_conversion import fen_to_tensor
import json
import torch.nn as nn

def test(path):
    model = CNN.CNN()
    model.load_state_dict(torch.load(path))
    model.eval()
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    model.to(device)

    with torch.no_grad():
        with open("Dataset/test_dataset") as file:
            total_actual = []
            total_pred = []
            for line in file:
                data = json.loads(line)
                fen = data["fen"]
                cp = data["cp"]
                tensor = fen_to_tensor(fen).to(device)
                tensor = tensor.unsqueeze(0)
                cp_final = torch.tensor([data["cp"] / 100.0], dtype=torch.float32).to(device)
                pred = model(tensor)

                total_pred.append(pred.item())
                total_actual.append(cp_final.item())
            MSE = nn.MSELoss()
            final_prediction = torch.tensor(total_pred)
            final_actual = torch.tensor(total_actual)
            loss = MSE(final_prediction, final_actual)
            return loss.item()