Strange execution time on CPU

Following code takes 32s for CPU to finish.

CPU times: user 1min 17s, sys: 386 ms, total: 1min 17s
Wall time: 32.1 s

By increasing EMBEDDING_DIM = 126 to EMBEDDING_DIM = 127,
it will take 5.19 s for CPU.

CPU times: user 8.91 s, sys: 2.62 s, total: 11.5 s
Wall time: 5.19 s

It supposed to take longer time with larger EMBEDDING_DIM, anyone knows what happened here?

import torch
import torch.autograd as autograd
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim

test_sentence = """When forty winters shall besiege thy brow,
And dig deep trenches in thy beauty's field,
Thy youth's proud livery so gazed on now,
Will be a totter'd weed of small worth held:
Then being asked, where all thy beauty lies,
Where all the treasure of thy lusty days;
To say, within thine own deep sunken eyes,
Were an all-eating shame, and thriftless praise.
How much more praise deserv'd thy beauty's use,
If thou couldst answer 'This fair child of mine
Shall sum my count, and make my old excuse,'
Proving his beauty by succession thine!
This were to be new made when thou art old,
And see thy blood warm when thou feel'st it cold.""".split()
trigrams = [([test_sentence[i], test_sentence[i + 1]], test_sentence[i + 2])
            for i in range(len(test_sentence) - 2)]

vocab = set(test_sentence)
word_to_ix = {word: i for i, word in enumerate(vocab)}

class NGramLanguageModeler(nn.Module):

    def __init__(self, vocab_size, embedding_dim, context_size):
        super(NGramLanguageModeler, self).__init__()
        self.embeddings = nn.Embedding(vocab_size, embedding_dim)
        self.linear1 = nn.Linear(context_size * embedding_dim, 128)
        self.linear2 = nn.Linear(128, vocab_size)

    def forward(self, inputs):
        embeds = self.embeddings(inputs).view((1, -1)) # here the inputs will be 10x2
        #since using two words to predict the third word
        out = F.relu(self.linear1(embeds))
        out = self.linear2(out)
        log_probs = F.log_softmax(out, dim=1)
        return log_probs

losses = []
loss_function = nn.NLLLoss()
model = NGramLanguageModeler(len(vocab), EMBEDDING_DIM, CONTEXT_SIZE)
optimizer = optim.SGD(model.parameters(), lr=0.001)

for epoch in range(100):
    total_loss = torch.Tensor([0])
    for context, target in trigrams:
        context_idxs = [word_to_ix[w] for w in context]
        context_var = autograd.Variable(torch.LongTensor(context_idxs))
        log_probs = model(context_var)
        loss = loss_function(log_probs, autograd.Variable(

        total_loss +=
print(losses)  # The loss decreased every iteration over the training data!