2022-05-06 20:20:22 +02:00
|
|
|
import matplotlib.pyplot as plt
|
2022-05-05 22:33:52 +02:00
|
|
|
import torch
|
|
|
|
from torch.utils.data import DataLoader
|
2022-05-06 20:20:22 +02:00
|
|
|
|
2022-05-06 21:51:49 +02:00
|
|
|
from model import MLP, PlantsDataset, test
|
2022-05-06 20:20:22 +02:00
|
|
|
|
2022-05-05 22:33:52 +02:00
|
|
|
|
|
|
|
def load_model():
|
|
|
|
model = MLP()
|
|
|
|
model.load_state_dict(torch.load('./model_out'))
|
|
|
|
return model
|
|
|
|
|
|
|
|
|
|
|
|
def load_dev_dataset(batch_size=64):
|
|
|
|
plant_dev = PlantsDataset('data/Plant_1_Generation_Data.csv.dev')
|
|
|
|
return DataLoader(plant_dev, batch_size=batch_size)
|
|
|
|
|
|
|
|
|
2022-05-05 22:51:30 +02:00
|
|
|
def make_plot(values):
|
|
|
|
build_nums = list(range(1, len(values) + 1))
|
|
|
|
plt.xlabel('Build number')
|
|
|
|
plt.ylabel('MSE Loss')
|
|
|
|
plt.plot(build_nums, values, label='Model MSE Loss over builds')
|
|
|
|
plt.legend()
|
|
|
|
plt.savefig('trend.png')
|
|
|
|
|
|
|
|
|
2022-05-05 22:33:52 +02:00
|
|
|
def main():
|
|
|
|
model = load_model()
|
|
|
|
dataloader = load_dev_dataset()
|
2022-05-06 20:20:22 +02:00
|
|
|
|
2022-05-05 22:33:52 +02:00
|
|
|
loss_fn = torch.nn.MSELoss()
|
|
|
|
|
2022-05-05 22:40:07 +02:00
|
|
|
loss = test(dataloader, model, loss_fn)
|
2022-05-05 23:22:10 +02:00
|
|
|
with open('evaluation_results.txt', 'a+') as f:
|
2022-05-05 22:40:07 +02:00
|
|
|
f.write(f'{str(loss)}\n')
|
2022-05-06 20:20:22 +02:00
|
|
|
with open('evaluation_results.txt', 'r') as f:
|
2022-05-05 22:51:30 +02:00
|
|
|
values = [float(line) for line in f.readlines() if line]
|
|
|
|
make_plot(values)
|
2022-05-05 22:33:52 +02:00
|
|
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
2022-05-06 20:20:22 +02:00
|
|
|
main()
|