Inquiry about implementation of mean_wQuantileLoss and m_sum_mean_wQuantileLoss

[hanlaoshi]

I am currently working on reproducing the mean_wQuantileLoss and m_sum_mean_wQuantileLoss metrics from a research paper. However, I have encountered significant discrepancies between my results and those obtained from the model.

I would appreciate any assistance in identifying potential errors in my implementation.

Here is a summary of my approach:
def quantile_loss(y_true, y_pred, q):
loss = 2 * np.sum(np.abs((y_pred - y_true) * ((y_true <= y_pred) - q)))
return loss
def quantile_loss(y_true, y_pred, q):
loss = 2 * np.sum(np.abs((y_pred - y_true) * ((y_true <= y_pred) - q)))
return loss
def mean_wQuantileLoss(forecasts, targets, quantiles=(np.arange(20)/20.0)[1:]):
B, T, D = targets.shape
loss = 0.0
for i in range(B):
for d in range(D):
for q in quantiles:
loss += quantile_loss(targets[i, :, d], forecasts[i, :, d], q) / np.sum(np.abs(targets))
return loss / (B * D * len(quantiles))

def m_sum_mean_wQuantileLoss(forecasts, targets, quantiles=(np.arange(20)/20.0)[1:]):#
B, T, D = targets.shape
forecasts = np.sum(forecasts, axis=-1)
targets = np.sum(targets, axis=-1)
total_loss = 0.0
for i in range(B):
loss = 0.0
for q in quantiles:
loss += quantile_loss(targets[i], forecasts[i], q) / np.sum(np.abs(targets))
total_loss += loss / len(quantiles)
return total_loss / B