orbit.StandardEvaluator | TensorFlow v2.16.1 (original) (raw)

Implements the standard functionality of AbstractEvaluator APIs.

orbit.StandardEvaluator(
    eval_dataset,
    options: Optional[orbit.StandardEvaluatorOptions] = None
)

This class structures evaluation roughly as follows:

state = eval_begin()
for _ in range(num_steps):
  step_outputs = eval_step(eval_iterator)
  state = eval_reduce(state, step_outputs)
return eval_end(state)

Calls to eval_begin and eval_end are always done in eager mode, while eval_step may be compiled with tf.function as determined by the options passed to __init__. eval_reduce is in eager mode ifuse_tf_while_loop=False in StandardEvaluatorOptions, but in graph mode ifuse_tf_while_loop=True.

This class does not support completely evaluating multiple different datasets (i.e., where every example of each dataset should be processed, as opposed to running for a fixed number of evaluation steps). A custom AbstractEvaluatoris recommended in this case.

Args
eval_dataset	A tf.nest-compatible structure of tf.data.Dataset orDistributedDataset. On TPUs, if users want to exaust the dataset without specifying number of eval steps, it is recommended to setdrop_remainder=False when batching the dataset, so the infrastructure can handle the last partial batch properly.
options	An orbit.StandardEvaluatorOptions instance.

Attributes
eval_dataset	The current evaluation dataset.
name	Returns the name of this module as passed or determined in the ctor.
name_scope	Returns a tf.name_scope instance for this class.
non_trainable_variables	Sequence of non-trainable variables owned by this module and its submodules.
submodules	Sequence of all sub-modules.Submodules are modules which are properties of this module, or found as properties of modules which are properties of this module (and so on). a = tf.Module() b = tf.Module() c = tf.Module() a.b = b b.c = c list(a.submodules) == [b, c] True list(b.submodules) == [c] True list(c.submodules) == [] True
trainable_variables	Sequence of trainable variables owned by this module and its submodules.
variables	Sequence of variables owned by this module and its submodules.

Methods

`create_eval_loop_fn`

View source

create_eval_loop_fn(
    has_state: bool
)

Creates an eval loop from the current step function and options.

Args
has_state	If the step function has state, state will be kept in the loop.

Returns
The eval loop function, i.e. wrapper of multiple eval steps.

`eval_begin`

View source

eval_begin() -> Any

Called once at the beginning of the evaluation.

This method is always called in eager mode, and is a good place to reset metrics that accumulate values over the course of evaluation.

Note that this method is called before dataset iterator creation.

Returns
A value to pass as the state argument to eval_reduce.

`eval_end`

View source

eval_end(
    *args
) -> Optional[runner.Output]

Called at the end of the evaluation.

Called once at the end of evaluation.

This method is always called in eager mode, and is a good place to get metric results. The value returned from this function will be returned as-is from the evaluate method implementation provided by StandardEvaluator.

Args
*args	The outputs from eval_reduce for the last eval step, if they are non-None (if they are None, nothing is passed).

Returns
The function may return a dictionary of Tensors, which will be written to logs and as TensorBoard summaries. It can also be a nested dictionary, yielding a hierarchy of summary directories.

`eval_reduce`

View source

eval_reduce(
    state: Optional[Any] = None, step_outputs: Optional[runner.Output] = None
) -> Any

A function to perform per-step reduction on the evaluation outputs.

This is useful for passing state throughout evaluation, especially in cases where maintaining or accumulating state is hard to accomplish usingtf.metrics.Metric or other tf.Variable-based approaches. For instance, it can be used to easily accumulate all per-example losses from the full evaluation for subsequent processing in eval_end().

Args
state	A state being maintained throughout the evaluation.
step_outputs	Outputs from the current evaluation step.

Returns
An output which is passed as the state argument to this function for the next step. After evaluation is finished, the output from last step will be passed to eval_end.

`eval_step`

View source

@abc.abstractmethod eval_step( iterator ) -> Any

Implements one step of evaluation.

What a "step" consists of is up to the implementer. When using distribution strategies, the call to this method takes place in the "cross-replica context" for generality, to allow e.g. multiple iterator dequeues and calls to strategy.run.

Note that if use_tf_function=True, all the code inside eval_step should be compatible with tf.function tracing (and in particular, any state modifications involving self should be avoided). In some cases, non-tf.function compatible code can be moved to eval_loop_begin,eval_reduce, or eval_loop_end, which always execute eagerly.

Args
iterator	A tf.nest-compatible structure of tf.data.Iterator orDistributedIterator.

Returns
An output which is passed as step_outputs argument into eval_reducefunction.

`evaluate`

View source

evaluate(
    num_steps: tf.Tensor
) -> Optional[runner.Output]

Implements num_steps steps of evaluation.

Args
num_steps	The number of evaluation steps to run. When this is -1, evaluation proceeds until a call to eval_step raises a StopIterationor tf.errors.OutOfRangeError.

Returns
The output of self.eval_end().

Raises
ValueError	If options.use_tf_while_loop is True and num_steps is unspecified.

`with_name_scope`

@classmethod with_name_scope( method )

Decorator to automatically enter the module name scope.

class MyModule(tf.Module): @tf.Module.with_name_scope def __call__(self, x): if not hasattr(self, 'w'): self.w = tf.Variable(tf.random.normal([x.shape[1], 3])) return tf.matmul(x, self.w)

Using the above module would produce tf.Variables and tf.Tensors whose names included the module name:

mod = MyModule() mod(tf.ones([1, 2])) <tf.Tensor: shape=(1, 3), dtype=float32, numpy=..., dtype=float32)> mod.w <tf.Variable 'my_module/Variable:0' shape=(2, 3) dtype=float32, numpy=..., dtype=float32)>

Args
method	The method to wrap.

Returns
The original method wrapped such that it enters the module's name scope.