Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

No Case: Added more logs for debugging sampling. #67

Open
wants to merge 1 commit into
base: releases/0.8.6
Choose a base branch
from
Open

No Case: Added more logs for debugging sampling. #67

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions python/ray/util/iter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -512,6 +512,10 @@ def base_iterator(timeout=None):
results.append(ray.get(f))
except (StopIteration, RuntimeError) as ex_i:
if was_cause_by_stop_iteration(ex_i):
logger.exception(
"Encountered an exception while extracting "
"the valid data from `futures`."
)
if self.is_infinite_sequence:
stoped_actors.append(a)
else:
Expand All @@ -520,11 +524,17 @@ def base_iterator(timeout=None):
raise ex_i

if results:
logger.info(
f"Gathered {len(results)} shards of batch data "
f"with {len(stoped_actors)} stopped actors."
)
yield results
elif self.is_infinite_sequence and len(stoped_actors) == len(active):
raise ex
logger.info(f"Kicking off {len(active)} new sampling tasks.")
futures = [a.par_iter_next.remote() for a in active]
else:
logger.exception("Encountered an non-StopIteration exception.")
raise ex

name = "{}.batch_across_shards()".format(self)
Expand Down
10 changes: 10 additions & 0 deletions rllib/execution/rollout_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -158,6 +158,11 @@ def __init__(self, min_batch_size: int):
self.buffer = []
self.count = 0
self.batch_start_time = None
# We would like to log per 100 received batch shards. Since each
# shard size varies, the batch count may not exactly hit the 100
# multiplier, and this variable is used for resetting the count for
# every 100 or more shards.
self._count_for_log = 0

def _on_fetch_start(self):
if self.batch_start_time is None:
Expand All @@ -168,6 +173,7 @@ def __call__(self, batch: SampleBatchType) -> List[SampleBatchType]:
self.buffer.append(batch)
self.count += batch.count
if self.count >= self.min_batch_size:
logger.info(f"Completed to gather a full batch with size {self.count}.")
if self.count > self.min_batch_size * 2:
logger.info("Collected more training samples than expected "
"(actual={}, expected={}). ".format(
Expand All @@ -181,7 +187,11 @@ def __call__(self, batch: SampleBatchType) -> List[SampleBatchType]:
self.batch_start_time = None
self.buffer = []
self.count = 0
self._count_for_log = 0
return [out]
if self.count >= 100 + self._count_for_log:
logger.info(f"Gathered the partial batch with size {self.count}.")
self._count_for_log = self.count
return []


Expand Down