d8f8ac953bc290739aa32a0ad277d53820d4473a,pyprob/modules.py,ProposalUniformDiscrete,loss,#ProposalUniformDiscrete#Any#Any#,77

Before Change

        for b in range(batch_size):
            value = samples[b].value[0]
            min = samples[b].distribution.prior_min
            l -= log_weights[b, int(value) - min] // Should we average this over dimensions? See http://pytorch.org/docs/nn.html//torch.nn.KLDivLoss
        return l

class ProposalNormal(nn.Module):

After Change

        log_weights = torch.log(proposal_output + util.epsilon)
        l = 0
        for b in range(batch_size):
            value = Variable(samples[b].value, requires_grad=False) // value is one-hot
            l -= torch.sum(log_weights[b] * value) // Should we average this over dimensions? See http://pytorch.org/docs/nn.html//torch.nn.KLDivLoss
        return l

class ProposalNormal(nn.Module):

Instances