utils.utils

utils.utils

Functions

Name	Description
broadcast_1d_array	Broadcast a 1D array of shape (N,) to shape (N, *additional_dims)
safe_reshape	A safe reshaping function with the following properties:

broadcast_1d_array

utils.utils.broadcast_1d_array(arr_1d, additional_dims=())

Broadcast a 1D array of shape (N,) to shape (N, *additional_dims) with a single reshape operation.

Parameters:

arr_1d : numpy.ndarray or jax.numpy.ndarray 1D input array of shape (N,) additional_dims : tuple Additional dimensions to broadcast to. Can be empty tuple () for no additional dimensions.

Returns:

numpy.ndarray or jax.numpy.ndarray Broadcasted array of shape (N, *additional_dims)

safe_reshape

utils.utils.safe_reshape(arr, new_shape, fill_value=jnp.nan)

A safe reshaping function with the following properties: - If new_shape has fewer elements than arr, raises an error - If new_shape has equal elements to arr, performs standard reshape - If new_shape requires more elements than arr, fills extra space with fill_value

Args: arr: JAX array to reshape new_shape: Tuple of integers specifying the new shape fill_value: Value to use for filling extra elements (default: jnp.nan)

Returns: Reshaped JAX array

# utils.utils { #tvboptim.utils.utils }

`utils.utils`



## Functions

| Name | Description |
| --- | --- |
| [broadcast_1d_array](#tvboptim.utils.utils.broadcast_1d_array) | Broadcast a 1D array of shape (N,) to shape (N, *additional_dims) |
| [safe_reshape](#tvboptim.utils.utils.safe_reshape) | A safe reshaping function with the following properties: |

### broadcast_1d_array { #tvboptim.utils.utils.broadcast_1d_array }

```python
utils.utils.broadcast_1d_array(arr_1d, additional_dims=())
```

Broadcast a 1D array of shape (N,) to shape (N, *additional_dims)
with a single reshape operation.

#### Parameters: {.doc-section .doc-section-parameters}

arr_1d : numpy.ndarray or jax.numpy.ndarray
    1D input array of shape (N,)
additional_dims : tuple
    Additional dimensions to broadcast to. Can be empty tuple () for no additional dimensions.

#### Returns: {.doc-section .doc-section-returns}

numpy.ndarray or jax.numpy.ndarray
    Broadcasted array of shape (N, *additional_dims)

### safe_reshape { #tvboptim.utils.utils.safe_reshape }

```python
utils.utils.safe_reshape(arr, new_shape, fill_value=jnp.nan)
```

A safe reshaping function with the following properties:
- If new_shape has fewer elements than arr, raises an error
- If new_shape has equal elements to arr, performs standard reshape
- If new_shape requires more elements than arr, fills extra space with fill_value

Args:
    arr: JAX array to reshape
    new_shape: Tuple of integers specifying the new shape
    fill_value: Value to use for filling extra elements (default: jnp.nan)

Returns:
    Reshaped JAX array