Batch inference of softmax does not sum to 1

1

I am working with REINFORCE algorithm with PyTorch. I noticed that the batch inference/predictions of my simple network with Softmax doesn’t sum to 1 (not even close to 1). I am attaching a minimum working code so that you can reproduce it. What am I missing here?

import numpy as np
import torch

obs_size = 9
HIDDEN_SIZE = 9
n_actions = 2

np.random.seed(0)

model = torch.nn.Sequential(
            torch.nn.Linear(obs_size, HIDDEN_SIZE),
            torch.nn.ReLU(),
            torch.nn.Linear(HIDDEN_SIZE, n_actions),
            torch.nn.Softmax(dim=0)
        )

state_transitions = np.random.rand(3, obs_size)

state_batch = torch.Tensor([s for s in state_transitions])
pred_batch = model(state_batch)  # WRONG PREDICTIONS!
print('wrong predictions:\n', *pred_batch.detach().numpy())
# [0.34072137 0.34721774] [0.30972624 0.30191955] [0.3495524 0.3508627]
# DOES NOT SUM TO 1 !!!

pred_batch = [model(s).detach().numpy() for s in state_batch]  # CORRECT PREDICTIONS
print('correct predictions:\n', *pred_batch)
# [0.5955179  0.40448207] [0.6574412  0.34255883] [0.624833   0.37516695]
# DOES SUM TO 1 AS EXPECTED
2

In case 1: model(state_batch)

  • The output dimension of torch.nn.Linear is (3,2) - 2 Dimensional

In case 2: [model(s).detach().numpy() for s in state_batch]

  • The output dimension of torch.nn.Linear is just (2) - 1 Dimensional

The torch.nn.Softmax(dim=0) does softmax at dim=0 which will work properly for Case 2 as there is only 1 dimension but in case of Case 1 there are 2 dimensions and that is the root cause for the issue.

Changing your code to torch.nn.Softmax(dim=1) should fix this issue for Case 1.

And if you still want the new changed code to work for Case 2, then do the following,
pred_batch = [model(s.unsqueeze(dim=0).detach().numpy() for s in state_batch]

1 Like