returnn.frontend.backend

Public backend-related functions.

returnn.frontend.backend.select_backend(name: str)

Select backend by name.

Parameters:

name – “torch”, “tf”, “returnn_layers_tf”, “numpy”

returnn.frontend.backend.get_selected_backend() → str | None

Returns:

the selected backend name, or None if not selected

returnn.frontend.backend.select_backend_torch()

Selects the PyTorch (low-level) backend.

returnn.frontend.backend.select_backend_returnn_layers_tf()

Selects the RETURNN layers backend (based on TF).

returnn.frontend.backend.is_executing_eagerly() → bool

Returns:

whether the current selected backend is executing eagerly

returnn.frontend.backend.is_backend_raw_tensor_dim_tag_independent() → bool

Returns:

whether raw tensors of the backend are independent of Dim (Usually yes, e.g. tf.Tensor or torch.Tensor, but the TF-layers backend is an exception.)