Tensorflow equivalent of Pytorch FloatTensor

How can I implement this line of code with just using tensorflow?

self.kernel = nn.Parameter(torch.FloatTensor(embedding_size, num_groups)) nn.init.normal_(self.kernel, std=0.01)

How can I implement this line of code with just using tensorflow?

self.kernel = nn.Parameter(torch.FloatTensor(embedding_size, num_groups)) nn.init.normal_(self.kernel, std=0.01)