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.

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

Added ability to set directory in pytorch lightning #581

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open
Changes from 1 commit
Commits
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
12 changes: 8 additions & 4 deletions python/keepsake/pl_callback.py
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
from copy import deepcopy
from typing import Optional, Dict, Tuple, Any
from pathlib import Path

import keepsake
from pytorch_lightning.callbacks.base import Callback
Expand Down Expand Up @@ -56,15 +57,18 @@ def __init__(
"""

super().__init__()
self.filepath = filepath
self.filepath = Path(filepath).resolve()
self.params = params
self.primary_metric = primary_metric
self.period = period
self.save_weights_only = save_weights_only
self.last_global_step_saved = -1

def on_pretrain_routine_start(self, trainer, pl_module):
self.experiment = keepsake.init(path=".", params=self.params)
self.experiment = keepsake.init(
path=str(self.filepath.parent),
fishbotics marked this conversation as resolved.
Show resolved Hide resolved
params=self.params,
)

def on_epoch_end(self, trainer, pl_module):
self._save_model(trainer, pl_module)
Expand All @@ -89,7 +93,7 @@ def _save_model(self, trainer, pl_module):
return

if self.filepath != None:
trainer.save_checkpoint(self.filepath, self.save_weights_only)
trainer.save_checkpoint(self.filepath.name, self.save_weights_only)

self.last_global_step_saved = global_step

Expand All @@ -99,7 +103,7 @@ def _save_model(self, trainer, pl_module):
metrics.update({"global_step": trainer.global_step})

self.experiment.checkpoint(
path=self.filepath,
path=self.filepath.name,
step=trainer.current_epoch,
metrics=metrics,
primary_metric=self.primary_metric,
Expand Down