python - GPyTorch Multi-class Classification ; What is num_features in "gpytorch.likelihoods.SoftmaxLikelihood"?

Question

I am trying to make a Multi-class classifier based on the notebook by changing "likelihoods.BernoulliLikelihood" into "likelihoods.SoftmaxLikelihood".

However, I cannot find the suitable value for parameter num_features. I tried different values but all of them give error. I appreciate it if you guide me on this issue.

Code:

import torch
import gpytorch

from gpytorch.models import AbstractVariationalGP
from gpytorch.variational import CholeskyVariationalDistribution
from gpytorch.variational import VariationalStrategy
from gpytorch.mlls.variational_elbo import VariationalELBO


"""
Data
"""

train_x = torch.linspace(0, 1, 10)
train_y = torch.tensor([1, 1, 0, -1, 0, 1, -1, -1, 1, 1])

num_classes = 3

num_features = 1



"""
Model
"""

class GPClassificationModel(AbstractVariationalGP):
    def __init__(self, train_x):
        variational_distribution = CholeskyVariationalDistribution(train_x.size(0))   
        variational_strategy = VariationalStrategy(self, train_x, variational_distribution)
        super(GPClassificationModel, self).__init__(variational_strategy)
        self.mean_module = gpytorch.means.ConstantMean()
        self.covar_module = gpytorch.kernels.ScaleKernel(gpytorch.kernels.RBFKernel())

    def forward(self, x):
        mean_x = self.mean_module(x)
        covar_x = self.covar_module(x)
        latent_pred = gpytorch.distributions.MultivariateNormal(mean_x, covar_x)
        return latent_pred   
        

# Initialize model and likelihood
model = GPClassificationModel(train_x)
likelihood = gpytorch.likelihoods.SoftmaxLikelihood(num_features = num_features, num_classes=num_classes)


"""
Train
"""

model.train()
likelihood.train()

optimizer = torch.optim.Adam(model.parameters(), lr=0.1)

# "Loss" for GPs - the marginal log likelihood
# train_y.numel() refers to the amount of training data
mll = VariationalELBO(likelihood, model, train_y.numel())

training_iter = 50
for i in range(training_iter):
    # Zero backpropped gradients from previous iteration
    optimizer.zero_grad()
    # Get predictive output
    output = model(train_x)
    # Calc loss and backprop gradients
    loss = -mll(output, train_y)
    loss.backward()
    print('Iter %d/%d - Loss: %.3f' % (i + 1, training_iter, loss.item()))
    optimizer.step()

标签： pythonmulticlass-classificationgpytorch