lightning/pytorch_lightning/core/saving.py

383 lines
14 KiB
Python

# Copyright The PyTorch Lightning team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import ast
import csv
import inspect
import os
from argparse import Namespace
from typing import Union, Dict, Any, Optional, Callable, MutableMapping
import fsspec
import torch
import yaml
from pytorch_lightning import _logger as log
from pytorch_lightning.utilities import rank_zero_warn, AttributeDict
from pytorch_lightning.utilities.cloud_io import load as pl_load
from pytorch_lightning.utilities.cloud_io import get_filesystem
PRIMITIVE_TYPES = (bool, int, float, str)
ALLOWED_CONFIG_TYPES = (AttributeDict, MutableMapping, Namespace)
try:
from omegaconf import OmegaConf
except ImportError:
OmegaConf = None
# the older shall be on the top
CHECKPOINT_PAST_HPARAMS_KEYS = (
'hparams',
'module_arguments', # used in 0.7.6
)
class ModelIO(object):
CHECKPOINT_HYPER_PARAMS_KEY = 'hyper_parameters'
CHECKPOINT_HYPER_PARAMS_NAME = 'hparams_name'
CHECKPOINT_HYPER_PARAMS_TYPE = 'hparams_type'
@classmethod
def load_from_checkpoint(
cls,
checkpoint_path: str,
*args,
map_location: Optional[Union[Dict[str, str], str, torch.device, int, Callable]] = None,
hparams_file: Optional[str] = None,
strict: bool = True,
**kwargs,
):
r"""
Primary way of loading a model from a checkpoint. When Lightning saves a checkpoint
it stores the arguments passed to `__init__` in the checkpoint under `module_arguments`
Any arguments specified through \*args and \*\*kwargs will override args stored in `hparams`.
Args:
checkpoint_path: Path to checkpoint. This can also be a URL.
args: Any positional args needed to init the model.
map_location:
If your checkpoint saved a GPU model and you now load on CPUs
or a different number of GPUs, use this to map to the new setup.
The behaviour is the same as in :func:`torch.load`.
hparams_file: Optional path to a .yaml file with hierarchical structure
as in this example::
drop_prob: 0.2
dataloader:
batch_size: 32
You most likely won't need this since Lightning will always save the hyperparameters
to the checkpoint.
However, if your checkpoint weights don't have the hyperparameters saved,
use this method to pass in a .yaml file with the hparams you'd like to use.
These will be converted into a :class:`~dict` and passed into your
:class:`LightningModule` for use.
If your model's `hparams` argument is :class:`~argparse.Namespace`
and .yaml file has hierarchical structure, you need to refactor your model to treat
`hparams` as :class:`~dict`.
strict: Whether to strictly enforce that the keys in :attr:`checkpoint_path` match the keys
returned by this module's state dict. Default: `True`.
hparam_overrides: A dictionary with keys to override in the hparams
kwargs: Any keyword args needed to init the model.
Return:
:class:`LightningModule` with loaded weights and hyperparameters (if available).
Example:
.. code-block:: python
# load weights without mapping ...
MyLightningModule.load_from_checkpoint('path/to/checkpoint.ckpt')
# or load weights mapping all weights from GPU 1 to GPU 0 ...
map_location = {'cuda:1':'cuda:0'}
MyLightningModule.load_from_checkpoint(
'path/to/checkpoint.ckpt',
map_location=map_location
)
# or load weights and hyperparameters from separate files.
MyLightningModule.load_from_checkpoint(
'path/to/checkpoint.ckpt',
hparams_file='/path/to/hparams_file.yaml'
)
# override some of the params with new values
MyLightningModule.load_from_checkpoint(
PATH,
num_layers=128,
pretrained_ckpt_path: NEW_PATH,
)
# predict
pretrained_model.eval()
pretrained_model.freeze()
y_hat = pretrained_model(x)
"""
if map_location is not None:
checkpoint = pl_load(checkpoint_path, map_location=map_location)
else:
checkpoint = pl_load(checkpoint_path, map_location=lambda storage, loc: storage)
if hparams_file is not None:
extension = hparams_file.split('.')[-1]
if extension.lower() in ('csv'):
hparams = load_hparams_from_tags_csv(hparams_file)
elif extension.lower() in ('yml', 'yaml'):
hparams = load_hparams_from_yaml(hparams_file)
else:
raise ValueError('.csv, .yml or .yaml is required for `hparams_file`')
hparams['on_gpu'] = False
# overwrite hparams by the given file
checkpoint[cls.CHECKPOINT_HYPER_PARAMS_KEY] = hparams
# for past checkpoint need to add the new key
if cls.CHECKPOINT_HYPER_PARAMS_KEY not in checkpoint:
checkpoint[cls.CHECKPOINT_HYPER_PARAMS_KEY] = {}
# override the hparams with values that were passed in
checkpoint[cls.CHECKPOINT_HYPER_PARAMS_KEY].update(kwargs)
model = cls._load_model_state(checkpoint, *args, strict=strict, **kwargs)
return model
@classmethod
def _load_model_state(cls, checkpoint: Dict[str, Any], *cls_args, strict: bool = True, **cls_kwargs):
cls_spec = inspect.getfullargspec(cls.__init__)
cls_init_args_name = inspect.signature(cls.__init__).parameters.keys()
# pass in the values we saved automatically
if cls.CHECKPOINT_HYPER_PARAMS_KEY in checkpoint:
model_args = {}
# add some back compatibility, the actual one shall be last
for hparam_key in CHECKPOINT_PAST_HPARAMS_KEYS + (cls.CHECKPOINT_HYPER_PARAMS_KEY,):
if hparam_key in checkpoint:
model_args.update(checkpoint[hparam_key])
model_args = _convert_loaded_hparams(model_args, checkpoint.get(cls.CHECKPOINT_HYPER_PARAMS_TYPE))
args_name = checkpoint.get(cls.CHECKPOINT_HYPER_PARAMS_NAME)
if args_name == 'kwargs':
# in case the class cannot take any extra argument filter only the possible
cls_kwargs.update(**model_args)
elif args_name:
if args_name in cls_init_args_name:
cls_kwargs.update({args_name: model_args})
else:
cls_args = (model_args,) + cls_args
if not cls_spec.varkw:
# filter kwargs according to class init unless it allows any argument via kwargs
cls_kwargs = {k: v for k, v in cls_kwargs.items() if k in cls_init_args_name}
# prevent passing positional arguments if class does not accept any
if len(cls_spec.args) <= 1 and not cls_spec.varargs and not cls_spec.kwonlyargs:
cls_args, cls_kwargs = [], {}
model = cls(*cls_args, **cls_kwargs)
# load the state_dict on the model automatically
model.load_state_dict(checkpoint['state_dict'], strict=strict)
# give model a chance to load something
model.on_load_checkpoint(checkpoint)
return model
def on_load_checkpoint(self, checkpoint: Dict[str, Any]) -> None:
"""
Do something with the checkpoint.
Gives model a chance to load something before ``state_dict`` is restored.
Args:
checkpoint: A dictionary with variables from the checkpoint.
"""
def on_save_checkpoint(self, checkpoint: Dict[str, Any]) -> None:
"""
Give the model a chance to add something to the checkpoint.
``state_dict`` is already there.
Args:
checkpoint: A dictionary in which you can save variables to save in a checkpoint.
Contents need to be pickleable.
"""
# -------------------------
# OPTIONAL HOOKS
# -------------------------
def on_hpc_save(self, checkpoint: Dict[str, Any]) -> None:
"""
Hook to do whatever you need right before Slurm manager saves the model.
Args:
checkpoint: A dictionary in which you can save variables to save in a checkpoint.
Contents need to be pickleable.
"""
def on_hpc_load(self, checkpoint: Dict[str, Any]) -> None:
"""
Hook to do whatever you need right before Slurm manager loads the model.
Args:
checkpoint: A dictionary with variables from the checkpoint.
"""
def _convert_loaded_hparams(model_args: dict, hparams_type: Union[Callable, str] = None) -> object:
"""Convert hparams according given type in callable or string (past) format"""
# if not hparams type define
if not hparams_type:
return model_args
# if past checkpoint loaded, convert str to callable
if isinstance(hparams_type, str):
hparams_type = AttributeDict
# convert hparams
return hparams_type(model_args)
def update_hparams(hparams: dict, updates: dict) -> None:
"""
Overrides hparams with new values
>>> hparams = {'c': 4}
>>> update_hparams(hparams, {'a': {'b': 2}, 'c': 1})
>>> hparams['a']['b'], hparams['c']
(2, 1)
>>> update_hparams(hparams, {'a': {'b': 4}, 'c': 7})
>>> hparams['a']['b'], hparams['c']
(4, 7)
Args:
hparams: the original params and also target object
updates: new params to be used as update
"""
for k, v in updates.items():
# if missing, add the key
if k not in hparams:
hparams[k] = v
continue
# recurse if dictionary
if isinstance(v, dict):
update_hparams(hparams[k], updates[k])
else:
# update the value
hparams.update({k: v})
def load_hparams_from_tags_csv(tags_csv: str) -> Dict[str, Any]:
"""Load hparams from a file.
>>> hparams = Namespace(batch_size=32, learning_rate=0.001, data_root='./any/path/here')
>>> path_csv = os.path.join('.', 'testing-hparams.csv')
>>> save_hparams_to_tags_csv(path_csv, hparams)
>>> hparams_new = load_hparams_from_tags_csv(path_csv)
>>> vars(hparams) == hparams_new
True
>>> os.remove(path_csv)
"""
fs = get_filesystem(tags_csv)
if not fs.exists(tags_csv):
rank_zero_warn(f"Missing Tags: {tags_csv}.", RuntimeWarning)
return {}
with fs.open(tags_csv, "r", newline="") as fp:
csv_reader = csv.reader(fp, delimiter=",")
tags = {row[0]: convert(row[1]) for row in list(csv_reader)[1:]}
return tags
def save_hparams_to_tags_csv(tags_csv: str, hparams: Union[dict, Namespace]) -> None:
fs = get_filesystem(tags_csv)
if not fs.isdir(os.path.dirname(tags_csv)):
raise RuntimeError(f"Missing folder: {os.path.dirname(tags_csv)}.")
if isinstance(hparams, Namespace):
hparams = vars(hparams)
with fs.open(tags_csv, "w", newline="") as fp:
fieldnames = ["key", "value"]
writer = csv.DictWriter(fp, fieldnames=fieldnames)
writer.writerow({"key": "key", "value": "value"})
for k, v in hparams.items():
writer.writerow({"key": k, "value": v})
def load_hparams_from_yaml(config_yaml: str) -> Dict[str, Any]:
"""Load hparams from a file.
>>> hparams = Namespace(batch_size=32, learning_rate=0.001, data_root='./any/path/here')
>>> path_yaml = './testing-hparams.yaml'
>>> save_hparams_to_yaml(path_yaml, hparams)
>>> hparams_new = load_hparams_from_yaml(path_yaml)
>>> vars(hparams) == hparams_new
True
>>> os.remove(path_yaml)
"""
fs = get_filesystem(config_yaml)
if not fs.exists(config_yaml):
rank_zero_warn(f"Missing Tags: {config_yaml}.", RuntimeWarning)
return {}
with fs.open(config_yaml, "r") as fp:
tags = yaml.full_load(fp)
return tags
def save_hparams_to_yaml(config_yaml, hparams: Union[dict, Namespace]) -> None:
"""
Args:
config_yaml: path to new YAML file
hparams: parameters to be saved
"""
fs = get_filesystem(config_yaml)
if not fs.isdir(os.path.dirname(config_yaml)):
raise RuntimeError(f"Missing folder: {os.path.dirname(config_yaml)}.")
# convert Namespace or AD to dict
if isinstance(hparams, Namespace):
hparams = vars(hparams)
elif isinstance(hparams, AttributeDict):
hparams = dict(hparams)
# saving with OmegaConf objects
if OmegaConf is not None:
if OmegaConf.is_config(hparams):
OmegaConf.save(hparams, config_yaml, resolve=True)
return
for v in hparams.values():
if OmegaConf.is_config(v):
OmegaConf.save(OmegaConf.create(hparams), config_yaml, resolve=True)
return
# saving the standard way
assert isinstance(hparams, dict)
with fs.open(config_yaml, "w", newline="") as fp:
yaml.dump(hparams, fp)
def convert(val: str) -> Union[int, float, bool, str]:
try:
return ast.literal_eval(val)
except (ValueError, SyntaxError) as err:
log.debug(err)
return val