VictorMylle/Thesis
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Added time as input feature

Browse Source
This commit is contained in:
Victor Mylle
2023-11-26 18:43:03 +00:00
parent a2c9427d16
commit 2f40f41786
9 changed files with 168 additions and 138 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
src/trainers/quantile_trainer.py
View File

@@ -218,7 +218,7 @@ class AutoRegressiveQuantileTrainer(AutoRegressiveTrainer):
if other_features is not None:
prev_features = torch.cat(
new_features, other_features, dim=1
(new_features.to(self.device), other_features.to(self.device)), dim=1
) # (batch_size, 96 + new_features)
else:
prev_features = new_features
@@ -252,36 +252,39 @@ class AutoRegressiveQuantileTrainer(AutoRegressiveTrainer):
def plot_quantile_percentages(
self, task, data_loader, train: bool = True, iteration: int = None
):
quantiles = self.quantiles.cpu().numpy()
total = 0
quantile_counter = {q: 0 for q in self.quantiles.cpu().numpy()}
quantile_counter = {q: 0 for q in quantiles}
self.model.eval()
with torch.no_grad():
for inputs, targets, _ in data_loader:
inputs = inputs.to("cuda")
output = self.model(inputs)
inputs = inputs.to(self.device)
output = self.model(inputs).cpu().numpy()
targets = targets.squeeze(-1).cpu().numpy()
# output shape: (batch_size, num_quantiles)
# target shape: (batch_size, 1)
for i, q in enumerate(self.quantiles.cpu().numpy()):
for i, q in enumerate(quantiles):
quantile_counter[q] += np.sum(
targets.squeeze(-1).cpu().numpy() < output[:, i].cpu().numpy()
targets < output[:, i]
)
total += len(targets)
# to numpy array of length len(quantiles)
percentages = np.array(
[quantile_counter[q] / total for q in self.quantiles.cpu().numpy()]
[quantile_counter[q] / total for q in quantiles]
)
bar_width = 0.35
index = np.arange(len(self.quantiles.cpu().numpy()))
index = np.arange(len(quantiles))
# Plotting the bars
fig, ax = plt.subplots(figsize=(15, 10))
bar1 = ax.bar(
index, self.quantiles.cpu().numpy(), bar_width, label="Ideal", color="brown"
index, quantiles, bar_width, label="Ideal", color="brown"
)
bar2 = ax.bar(
index + bar_width, percentages, bar_width, label="NN model", color="blue"
@@ -305,7 +308,7 @@ class AutoRegressiveQuantileTrainer(AutoRegressiveTrainer):
ax.set_ylabel("Fraction of data under quantile forecast")
ax.set_title(f"Quantile Performance Comparison ({series_name})")
ax.set_xticks(index + bar_width / 2)
ax.set_xticklabels(self.quantiles.cpu().numpy())
ax.set_xticklabels(quantiles)
ax.legend()
task.get_logger().report_matplotlib_figure(