olympus.tasks.task module

class olympus.tasks.task.GenerateSummary

Bases: object

Methods

get_name
is_nested
print
rename
retrieve_nested
task_summary

dispatch = {'DataLoader': <function GenerateSummary.<lambda> at 0x7f68e0704b70>, 'LRSchedule': <function GenerateSummary.<lambda> at 0x7f68e0704d08>, 'MetricList': <function GenerateSummary.<lambda> at 0x7f68e0704c80>, 'Model': <function GenerateSummary.<lambda> at 0x7f68e0704a60>, 'Optimizer': <function GenerateSummary.<lambda> at 0x7f68e0704ae8>, 'TransformedSubset': <function GenerateSummary.<lambda> at 0x7f68e0704bf8>}

get_name(self, attr, obj, type_name, depth=0)

is_nested(self, name)

print(self, msg='', end='n')

rename(self, name)

retrieve_nested(self, name, obj)

task_summary(self, obj)

class olympus.tasks.task.Task(device=None)

Bases: object

Attributes:

device

metrics

Methods

eval_loss(self, batch)	This is used to compute validation and test loss
fit(self, epoch[, context])	Execute a single batch
get_space(self, \*\*fidelities)	Return missing hyper parameters that need to be set using init
init(self, \*\*kwargs)	Used to initialize the hyperparameters is any
load_state_dict(self, state[, strict])	Try to load a previous unfinished state to resume
state_dict(self[, destination, prefix, …])	Save a state the task can go back to if an error occur

report
resumed
set_device
summary

device

eval_loss(self, batch)

This is used to compute validation and test loss

fit(self, epoch, context=None)

Execute a single batch

Parameters:

epoch: int

current step in the training process

context: dict

Optional Context

Notes

You should wrap whatever code you have here inside a BadResumeGuard to prevent users from resuming a failed task that can have a bad states

To resume a task, you need to create a clean one with the same hyper parameters. It will pickup automatically where at its last checkpoint

get_space(self, **fidelities)

Return missing hyper parameters that need to be set using init

init(self, **kwargs)

Used to initialize the hyperparameters is any

load_state_dict(self, state, strict=True)

Try to load a previous unfinished state to resume

Notes

You should wrap whatever code you have here inside a BadResumeGuard to prevent users from resuming a failed task that can have a bad states

To resume a task, you need to create a clean one with the same hyper parameters. It will pickup automatically where at its last checkpoint

metrics

report(self, pprint=True, print_fun=<built-in function print>)

resumed(self)

set_device(self, device)

state_dict(self, destination=None, prefix='', keep_vars=False)

Save a state the task can go back to if an error occur

summary(self)