nshtrainer 0.16.1__tar.gz → 0.17.1__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/PKG-INFO +1 -1
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/pyproject.toml +1 -1
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/_checkpoint/metadata.py +14 -12
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/_environment_info.py +6 -6
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/README.md +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/_checkpoint/loader.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/_checkpoint/saver.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/_experimental/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/_throughput_monitor_callback.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/actsave.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/base.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/_base.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/best_checkpoint.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/last_checkpoint.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/on_exception_checkpoint.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/early_stopping.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/ema.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/finite_checks.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/gradient_skipping.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/interval.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/log_epoch.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/norm_logging.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/print_table.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/throughput_monitor.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/timer.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/wandb_watch.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/data/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/data/balanced_batch_sampler.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/data/transform.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/_experimental.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/actsave.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/callbacks.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/config.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/data.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/log.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/lr_scheduler.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/model.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/nn.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/optimizer.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/runner.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/snapshot.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/snoop.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/trainer.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/typecheck.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/ll/util.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/loggers/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/loggers/_base.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/loggers/csv.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/loggers/tensorboard.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/loggers/wandb.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/lr_scheduler/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/lr_scheduler/_base.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/lr_scheduler/linear_warmup_cosine.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/lr_scheduler/reduce_lr_on_plateau.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/metrics/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/metrics/_config.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/base.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/config.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/callback.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/debug.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/distributed.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/logger.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/profiler.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/rlp_sanity_checks.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/model/modules/shared_parameters.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/nn/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/nn/mlp.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/nn/module_dict.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/nn/module_list.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/nn/nonlinearity.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/optimizer.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/runner.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/scripts/find_packages.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/trainer/__init__.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/trainer/_runtime_callback.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/trainer/checkpoint_connector.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/trainer/signal_connector.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/trainer/trainer.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/_useful_types.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/environment.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/seed.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/slurm.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/typed.py +0 -0
- {nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/util/typing_utils.py +0 -0
|
@@ -10,6 +10,8 @@ import nshconfig as C
|
|
|
10
10
|
import numpy as np
|
|
11
11
|
import torch
|
|
12
12
|
|
|
13
|
+
from ..util._environment_info import EnvironmentConfig
|
|
14
|
+
|
|
13
15
|
if TYPE_CHECKING:
|
|
14
16
|
from ..model import BaseConfig, LightningModuleBase
|
|
15
17
|
from ..trainer.trainer import Trainer
|
|
@@ -36,13 +38,13 @@ class CheckpointMetadata(C.Config):
|
|
|
36
38
|
global_step: int
|
|
37
39
|
training_time: datetime.timedelta
|
|
38
40
|
metrics: dict[str, Any]
|
|
39
|
-
environment:
|
|
41
|
+
environment: EnvironmentConfig
|
|
40
42
|
|
|
41
|
-
hparams:
|
|
43
|
+
hparams: Any
|
|
42
44
|
|
|
43
45
|
@classmethod
|
|
44
46
|
def from_file(cls, path: Path):
|
|
45
|
-
return cls.model_validate_json(path.read_text())
|
|
47
|
+
return cls.model_validate_json(path.read_text(encoding="utf-8"))
|
|
46
48
|
|
|
47
49
|
@classmethod
|
|
48
50
|
def from_ckpt_path(cls, checkpoint_path: Path):
|
|
@@ -89,8 +91,8 @@ def _generate_checkpoint_metadata(
|
|
|
89
91
|
global_step=trainer.global_step,
|
|
90
92
|
training_time=training_time,
|
|
91
93
|
metrics=metrics,
|
|
92
|
-
environment=config.environment
|
|
93
|
-
hparams=config.model_dump(
|
|
94
|
+
environment=config.environment,
|
|
95
|
+
hparams=config.model_dump(),
|
|
94
96
|
)
|
|
95
97
|
|
|
96
98
|
|
|
@@ -107,9 +109,9 @@ def _write_checkpoint_metadata(
|
|
|
107
109
|
|
|
108
110
|
# Write the metadata to the checkpoint directory
|
|
109
111
|
try:
|
|
110
|
-
metadata_path.write_text(metadata.model_dump_json(indent=4))
|
|
111
|
-
except Exception
|
|
112
|
-
log.
|
|
112
|
+
metadata_path.write_text(metadata.model_dump_json(indent=4), encoding="utf-8")
|
|
113
|
+
except Exception:
|
|
114
|
+
log.exception(f"Failed to write metadata to {checkpoint_path}")
|
|
113
115
|
else:
|
|
114
116
|
log.debug(f"Checkpoint metadata written to {checkpoint_path}")
|
|
115
117
|
|
|
@@ -118,8 +120,8 @@ def _remove_checkpoint_metadata(checkpoint_path: Path):
|
|
|
118
120
|
path = checkpoint_path.with_suffix(CheckpointMetadata.PATH_SUFFIX)
|
|
119
121
|
try:
|
|
120
122
|
path.unlink(missing_ok=True)
|
|
121
|
-
except Exception
|
|
122
|
-
log.
|
|
123
|
+
except Exception:
|
|
124
|
+
log.exception(f"Failed to remove {path}")
|
|
123
125
|
else:
|
|
124
126
|
log.debug(f"Removed {path}")
|
|
125
127
|
|
|
@@ -142,8 +144,8 @@ def _link_checkpoint_metadata(checkpoint_path: Path, linked_checkpoint_path: Pat
|
|
|
142
144
|
# on Windows, special permissions are required to create symbolic links as a regular user
|
|
143
145
|
# fall back to copying the file
|
|
144
146
|
shutil.copy(path, linked_path)
|
|
145
|
-
except Exception
|
|
146
|
-
log.
|
|
147
|
+
except Exception:
|
|
148
|
+
log.exception(f"Failed to link {path} to {linked_path}")
|
|
147
149
|
else:
|
|
148
150
|
log.debug(f"Linked {path} to {linked_path}")
|
|
149
151
|
|
|
@@ -429,12 +429,12 @@ class EnvironmentPackageConfig(C.Config):
|
|
|
429
429
|
version=clean_version,
|
|
430
430
|
path=Path(str(f)) if (f := dist.locate_file("")) else None,
|
|
431
431
|
summary=metadata["Summary"] if "Summary" in metadata else None,
|
|
432
|
-
author=metadata["Author"] if "
|
|
433
|
-
license=metadata["License"] if "
|
|
432
|
+
author=metadata["Author"] if "Author" in metadata else None,
|
|
433
|
+
license=metadata["License"] if "License" in metadata else None,
|
|
434
434
|
requires=requires,
|
|
435
435
|
)
|
|
436
|
-
except Exception
|
|
437
|
-
log.
|
|
436
|
+
except Exception:
|
|
437
|
+
log.exception(f"Error processing package {dist.name}")
|
|
438
438
|
|
|
439
439
|
except ImportError:
|
|
440
440
|
log.warning(
|
|
@@ -672,8 +672,8 @@ class GitRepositoryConfig(C.Config):
|
|
|
672
672
|
draft.is_dirty = repo.is_dirty()
|
|
673
673
|
except git.InvalidGitRepositoryError:
|
|
674
674
|
draft.is_git_repo = False
|
|
675
|
-
except Exception
|
|
676
|
-
log.
|
|
675
|
+
except Exception:
|
|
676
|
+
log.exception("Failed to get Git repository information")
|
|
677
677
|
draft.is_git_repo = None
|
|
678
678
|
|
|
679
679
|
return draft.finalize()
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/_throughput_monitor_callback.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/best_checkpoint.py
RENAMED
|
File without changes
|
{nshtrainer-0.16.1 → nshtrainer-0.17.1}/src/nshtrainer/callbacks/checkpoint/last_checkpoint.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|