tests/core/test_exp_manager.py

# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
#
# 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 json
import math
import os
import re
from pathlib import Path
from typing import Any

import lightning.pytorch as pl
import pytest
import torch
from lightning.pytorch import Callback
from lightning.pytorch.loops import _TrainingEpochLoop
from omegaconf import OmegaConf
from omegaconf.errors import OmegaConfBaseException

from nemo.collections.nlp.parts.nlp_overrides import NLPDDPStrategy
from nemo.constants import NEMO_ENV_VARNAME_VERSION
from nemo.core.classes import ModelPT
from nemo.utils.app_state import AppState
from nemo.utils.callbacks import NeMoModelCheckpoint
from nemo.utils.exp_manager import (
    CheckpointMisconfigurationError,
    LoggerMisconfigurationError,
    NotFoundError,
    exp_manager,
)


class MyTestOptimizer(torch.optim.Optimizer):
    def __init__(self, params):
        self._step = 0
        super().__init__(params, {})

    @torch.no_grad()
    def step(self, closure=None):
        loss = None
        if closure is not None:
            with torch.enable_grad():
                loss = closure()
        for group in self.param_groups:
            for p in group['params']:
                if self._step == 0:
                    p.data = 0.1 * torch.ones(p.shape)
                elif self._step == 1:
                    p.data = 0.0 * torch.ones(p.shape)
                else:
                    p.data = 0.01 * torch.ones(p.shape)
        self._step += 1
        return loss


class DoNothingOptimizer(torch.optim.Optimizer):
    def __init__(self, params):
        self._step = 0
        super().__init__(params, {})

    @torch.no_grad()
    def step(self, closure=None):
        loss = None
        if closure is not None:
            with torch.enable_grad():
                loss = closure()
        self._step += 1
        return loss


class OnesDataset(torch.utils.data.Dataset):
    def __init__(self, dataset_len):
        super().__init__()
        self.__dataset_len = dataset_len

    def __getitem__(self, *args):
        return torch.ones(2)

    def __len__(self):
        return self.__dataset_len


class ExampleModel(ModelPT):
    def __init__(self, *args, **kwargs):
        cfg = OmegaConf.structured({})
        super().__init__(cfg)
        pl.seed_everything(1234)
        self.l1 = torch.nn.modules.Linear(in_features=2, out_features=1)

    def train_dataloader(self):
        dataset = OnesDataset(2)
        return torch.utils.data.DataLoader(dataset, batch_size=2, num_workers=8)

    def val_dataloader(self):
        dataset = OnesDataset(10)
        return torch.utils.data.DataLoader(dataset, batch_size=2, num_workers=8)

    def forward(self, batch):
        output = self.l1(batch)
        output = torch.nn.functional.l1_loss(output, torch.zeros(output.size()).to(output.device))
        return output

    def validation_step(self, batch, batch_idx):
        self.loss = self(batch)
        return self.loss

    def training_step(self, batch, batch_idx):
        return self(batch)

    def configure_optimizers(self):
        return MyTestOptimizer(self.parameters())
        # return torch.optim.Adam(self.parameters(), lr=0.1)

    def list_available_models(self):
        pass

    def setup_training_data(self):
        pass

    def setup_validation_data(self):
        pass

    def on_validation_epoch_end(self):
        self.log("val_loss", torch.stack([self.loss]).mean())


class ExampleMCoreModel(ExampleModel):
    def sharded_state_dict(self):
        return {'a': 3}


class DoNothingModel(ExampleModel):
    def configure_optimizers(self):
        return DoNothingOptimizer(self.parameters())


class TestExpManager:
    @pytest.mark.unit
    def test_omegaconf(self):
        """Ensure omegaconf raises an error when an unexcepted argument is passed"""
        with pytest.raises(OmegaConfBaseException):
            exp_manager(pl.Trainer(accelerator='cpu'), {"unused": 1})

    @pytest.mark.unit
    def test_trainer_loggers(self, tmp_path):
        """Test that a trainer with logger errors out with a number of arguments. Test that it works with
        create_tensorboard_logger set to False
        """
        test_trainer = pl.Trainer(accelerator='cpu')  # Should create logger and modelcheckpoint

        with pytest.raises(LoggerMisconfigurationError):  # Fails because exp_manager defaults to trainer
            exp_manager(test_trainer, {"exp_dir": str(tmp_path)})
        with pytest.raises(LoggerMisconfigurationError):  # Fails because exp_manager defaults to trainer
            exp_manager(test_trainer, {"explicit_log_dir": str(tmp_path)})
        with pytest.raises(LoggerMisconfigurationError):  # Fails because exp_manager defaults to trainer
            exp_manager(test_trainer, {"resume_if_exists": True})

        # Check that exp_manager uses trainer.logger, it's exp_dir, name, and version
        log_dir = exp_manager(test_trainer, {"create_tensorboard_logger": False, "create_checkpoint_callback": False})
        assert log_dir.resolve() == Path("./lightning_logs/version_0").resolve()
        assert Path("./lightning_logs").exists()
        assert Path("./lightning_logs/version_0").exists()

        # Check that a trainer without a logger gets a logger attached to it
        test_trainer = pl.Trainer(accelerator='cpu', logger=False)
        log_dir = exp_manager(
            test_trainer,
            {"create_tensorboard_logger": True, "create_checkpoint_callback": False, "exp_dir": str(tmp_path)},
        )
        assert isinstance(test_trainer.logger, pl.loggers.TensorBoardLogger)

        test_trainer = pl.Trainer(accelerator='cpu', logger=False)
        # Check that a create_wandb_logger=True errors out unless wandb_logger_kwargs is passed.
        with pytest.raises(ValueError):
            log_dir = exp_manager(
                test_trainer,
                {
                    "create_tensorboard_logger": False,
                    "create_checkpoint_callback": False,
                    "exp_dir": str(tmp_path),
                    "create_wandb_logger": True,
                },
            )
        # Check that a WandbLogger is attached to logger if create_wandb_logger=True and wandb_logger_kwargs has name
        # and project
        log_dir = exp_manager(
            test_trainer,
            {
                "create_tensorboard_logger": False,
                "create_checkpoint_callback": False,
                "exp_dir": str(tmp_path),
                "create_wandb_logger": True,
                "wandb_logger_kwargs": {"name": "", "project": "", "offline": True},
            },
        )
        assert isinstance(test_trainer.logger, pl.loggers.WandbLogger)

    @pytest.mark.unit
    def test_trainer_neptune_logger(self, tmp_path):
        pytest.importorskip("neptune", reason="could not import `neptune`, use `pip install neptune` to run this test")

        test_trainer = pl.Trainer(accelerator='cpu', logger=False)
        # Check that a create_neptune_logger=True errors out unless neptune_logger_kwargs is passed.
        with pytest.raises(ValueError):
            _ = exp_manager(
                test_trainer,
                {
                    "create_tensorboard_logger": False,
                    "create_checkpoint_callback": False,
                    "exp_dir": str(tmp_path),
                    "create_neptune_logger": True,
                },
            )
        # Check that a NeptuneLogger is attached to logger if create_neptune_logger=True and neptune_logger_kwargs has name
        # and project
        _ = exp_manager(
            test_trainer,
            {
                "create_tensorboard_logger": False,
                "create_checkpoint_callback": False,
                "exp_dir": str(tmp_path),
                "create_neptune_logger": True,
                "neptune_logger_kwargs": {"name": "", "project": "", "api_key": ""},
            },
        )
        assert isinstance(test_trainer.logger, pl.loggers.NeptuneLogger)

    @pytest.mark.unit
    def test_checkpoint_configurations(self):
        """Test that trainer creating modelcheckpoint and asking exp_manager to do it too results in errors, but
        is error free if only one is asked to do so.
        """
        disable_tb_logger = {"create_tensorboard_logger": False}
        test_trainer = pl.Trainer(accelerator='cpu')  # Should create logger and modelcheckpoint
        with pytest.raises(CheckpointMisconfigurationError):  # Fails because both try to create modelcheckpoint
            exp_manager(test_trainer, disable_tb_logger)

        # Should succeed without error
        exp_manager(test_trainer, {"create_checkpoint_callback": False, "create_tensorboard_logger": False})

        test_trainer_2 = pl.Trainer(accelerator='cpu', enable_checkpointing=False)
        exp_manager(test_trainer_2, disable_tb_logger)  # Should succeed without error

    @pytest.mark.unit
    def test_default_log_dir(self):
        """Check the default of ./nemo_experiments/default/datetime works as intended"""
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)

        log_dir = exp_manager(test_trainer, {"create_tensorboard_logger": False, "create_checkpoint_callback": False})
        assert (log_dir / "..").resolve() == Path("./nemo_experiments/default/").resolve()
        assert Path("./nemo_experiments").exists()
        assert Path("./nemo_experiments/default/").exists()
        sub_dirs = [x for x in Path("./nemo_experiments/default/").iterdir() if x.is_dir()]
        assert len(sub_dirs) == 1
        assert re.match(r"[0-9]{4}-[0-9]{2}-[0-9]{2}_[0-9]{2}-[0-9]{2}-[0-9]{2}", sub_dirs[0].name)

    @pytest.mark.unit
    def test_log_dir_overrides(self, monkeypatch, tmp_path):
        """Check a variety of trainer options with exp_manager"""
        # Checks that explicit_log_dir ignores exp_dir, name, and version
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        log_dir = exp_manager(test_trainer, {"explicit_log_dir": str(tmp_path / "test_log_dir_overrides")})
        assert log_dir.resolve() == (tmp_path / "test_log_dir_overrides").resolve()
        assert Path(tmp_path).exists()
        assert Path(tmp_path / "test_log_dir_overrides").exists()

        # Checks that exp_manager uses exp_dir, default name, and explicit version
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        log_dir = exp_manager(test_trainer, {"exp_dir": str(tmp_path / "test_no_name"), "version": 957})
        assert log_dir.resolve() == (tmp_path / "test_no_name" / "default" / "957").resolve()
        assert Path(tmp_path).exists()
        assert Path(tmp_path / "test_no_name" / "default" / "957").exists()

        monkeypatch.delenv(NEMO_ENV_VARNAME_VERSION, raising=False)
        # Checks that use_datetime_version False toggle works
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        log_dir = exp_manager(test_trainer, {"exp_dir": str(tmp_path / "test_no_name"), "use_datetime_version": False})
        assert log_dir.resolve() == (tmp_path / "test_no_name" / "default" / "version_0").resolve()
        assert Path(tmp_path).exists()
        assert Path(tmp_path / "test_no_name" / "default" / "version_0").exists()

        monkeypatch.delenv(NEMO_ENV_VARNAME_VERSION, raising=False)
        # Checks that use_datetime_version False toggle works and version increments
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        log_dir = exp_manager(test_trainer, {"exp_dir": str(tmp_path / "test_no_name"), "use_datetime_version": False})
        assert log_dir.resolve() == (tmp_path / "test_no_name" / "default" / "version_1").resolve()
        assert Path(tmp_path).exists()
        assert Path(tmp_path / "test_no_name" / "default" / "version_1").exists()

    @pytest.mark.unit
    def test_resume(self, tmp_path):
        """Tests the resume capabilities of exp_manager"""
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)

        # Error because explicit_log_dir does not exist
        with pytest.raises(NotFoundError):
            exp_manager(
                test_trainer,
                {
                    "exp_dir": str(tmp_path / "test_resume"),
                    "resume_if_exists": True,
                    "explicit_log_dir": "Does_not_exist",
                },
            )

        # Error because checkpoints folder does not exist
        with pytest.raises(NotFoundError):
            exp_manager(test_trainer, {"resume_if_exists": True, "exp_dir": str(tmp_path / "test_resume")})

        # No error because we tell exp_manager to ignore notfounderror
        exp_manager(
            test_trainer,
            {
                "resume_if_exists": True,
                "exp_dir": str(tmp_path / "test_resume_2"),
                "resume_ignore_no_checkpoint": True,
            },
        )

        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints").mkdir(parents=True)
        # Error because checkpoints do not exist in folder
        with pytest.raises(NotFoundError):
            exp_manager(
                test_trainer,
                {
                    "resume_if_exists": True,
                    "explicit_log_dir": str(tmp_path / "test_resume" / "default" / "version_0"),
                },
            )

        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel--end.ckpt").touch()
        # Error because *end.ckpt is in folder indicating that training has already finished
        with pytest.raises(ValueError):
            exp_manager(
                test_trainer,
                {
                    "resume_if_exists": True,
                    "explicit_log_dir": str(tmp_path / "test_resume" / "default" / "version_0"),
                },
            )

        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel--end.ckpt").unlink()
        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel--last.ckpt").touch()
        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel2--last.ckpt").touch()
        # Error because multiple *last.ckpt is in folder. If more than one, don't know which to restore
        with pytest.raises(ValueError):
            exp_manager(
                test_trainer,
                {
                    "resume_if_exists": True,
                    "explicit_log_dir": str(tmp_path / "test_resume" / "default" / "version_0"),
                },
            )

        # Finally succeed
        Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel2--last.ckpt").unlink()
        log_dir = exp_manager(
            test_trainer,
            {"resume_if_exists": True, "explicit_log_dir": str(tmp_path / "test_resume" / "default" / "version_0")},
        )
        checkpoint = Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel--last.ckpt")
        assert Path(test_trainer.ckpt_path).resolve() == checkpoint.resolve()

        # Succeed again and make sure that run_0 exists and previous log files were moved
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        exp_manager(test_trainer, {"resume_if_exists": True, "explicit_log_dir": str(log_dir)})
        checkpoint = Path(tmp_path / "test_resume" / "default" / "version_0" / "checkpoints" / "mymodel--last.ckpt")
        assert Path(test_trainer.ckpt_path).resolve() == checkpoint.resolve()
        prev_run_dir = Path(tmp_path / "test_resume" / "default" / "version_0" / "run_0")
        assert prev_run_dir.exists()
        prev_log = Path(tmp_path / "test_resume" / "default" / "version_0" / "run_0" / "lightning_logs.txt")
        assert prev_log.exists()

        # Error becasue `dirpath` specified and has no checkpoint
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        dirpath_checkpoint_dir = Path(tmp_path / "test_resume" / "dirpath_test" / "ckpts")
        dirpath_checkpoint_dir.mkdir(parents=True)
        with pytest.raises(NotFoundError):
            exp_manager(
                test_trainer,
                {
                    "resume_if_exists": True,
                    "checkpoint_callback_params": {"dirpath": str(dirpath_checkpoint_dir)},
                    "explicit_log_dir": str(log_dir),
                },
            )

        # Check that model loads from `dirpath` and not <log_dir>/checkpoints
        dirpath_log_dir = Path(tmp_path / "test_resume" / "dirpath_test" / "logs")
        dirpath_log_dir.mkdir(parents=True)
        dirpath_checkpoint = Path(dirpath_checkpoint_dir / "mymodel--last.ckpt")
        dirpath_checkpoint.touch()
        exp_manager(
            test_trainer,
            {
                "resume_if_exists": True,
                "checkpoint_callback_params": {"dirpath": str(dirpath_checkpoint_dir)},
                "explicit_log_dir": str(dirpath_log_dir),
            },
        )
        assert Path(test_trainer.ckpt_path).resolve() == dirpath_checkpoint.resolve()

    @pytest.mark.unit
    def test_nemo_checkpoint_save_best_model_1(self, tmp_path):
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4)
        exp_manager(
            test_trainer,
            {"checkpoint_callback_params": {"save_best_model": True}, "explicit_log_dir": str(tmp_path / "test")},
        )
        model = ExampleModel()
        test_trainer.fit(model)

        assert Path(str(tmp_path / "test" / "checkpoints" / "default.nemo")).exists()

        model = ExampleModel.restore_from(str(tmp_path / "test" / "checkpoints" / "default.nemo"))
        assert float(model(torch.tensor([1.0, 1.0], device=model.device))) == 0.0

    @pytest.mark.unit
    def test_nemo_checkpoint_save_best_model_2(self, tmp_path):
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4)
        exp_manager(
            test_trainer,
            {"explicit_log_dir": str(tmp_path / "test")},
        )
        model = ExampleModel()
        test_trainer.fit(model)

        assert Path(str(tmp_path / "test" / "checkpoints" / "default.nemo")).exists()

        model = ExampleModel.restore_from(str(tmp_path / "test" / "checkpoints" / "default.nemo"))
        assert math.fabs(float(model(torch.tensor([1.0, 1.0], device=model.device))) - 0.03) < 1e-5

    @pytest.mark.unit
    def test_nemo_checkpoint_always_save_nemo(self, tmp_path):
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_best_model": True, "always_save_nemo": True},
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer.fit(model)

        assert Path(str(tmp_path / "test" / "checkpoints" / "default.nemo")).exists()

        model = ExampleModel.restore_from(str(tmp_path / "test" / "checkpoints" / "default.nemo"))
        assert float(model(torch.tensor([1.0, 1.0], device=model.device))) == 0.0

    @pytest.mark.unit
    def test_nemo_checkpoint_doesnt_produce_too_many_nemo_ckpts(self, tmp_path):
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_best_model": True, "always_save_nemo": True, "save_top_k": 2},
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer.fit(model)

        assert Path(str(tmp_path / "test" / "checkpoints" / "default.nemo")).exists()
        assert (
            len(list((tmp_path / "test" / "checkpoints").glob("default*.nemo"))) == 1
        )  # check number of `.nemo` checkpoints

        model = ExampleModel.restore_from(str(tmp_path / "test" / "checkpoints" / "default.nemo"))
        assert float(model(torch.tensor([1.0, 1.0], device=model.device))) == 0.0

    @pytest.mark.unit
    def test_nemo_checkpoint_make_checkpoint_dir(self, tmp_path):
        test_trainer = pl.Trainer(
            accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4, check_val_every_n_epoch=5
        )
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_best_model": True, "always_save_nemo": True},
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer.fit(model)

        assert Path(str(tmp_path / "test" / "checkpoints" / "default.nemo")).exists()

    @pytest.mark.unit
    def test_nemo_checkpoint_restore_model(self, tmp_path):
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_top_k": 1, "save_last": True},
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer.fit(model)

        checkpoint = list(Path(str(tmp_path / "test" / "checkpoints")).glob("*.ckpt"))
        # Make sure that only the best and last checkpoint is saved
        assert len(checkpoint) == 2
        assert math.fabs(float(model(torch.tensor([1.0, 1.0], device=model.device))) - 0.03) < 1e-5

        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=5)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_top_k": 1, "save_last": False},
                "explicit_log_dir": str(tmp_path / "test"),
                "resume_if_exists": True,
                "resume_past_end": True,
            },
        )
        model = DoNothingModel()
        model.l1.weight = torch.nn.Parameter(torch.tensor((0.0, 0.0)).unsqueeze(0))
        model.l1.bias = torch.nn.Parameter(torch.tensor(1.0))
        assert math.fabs(float(model(torch.tensor([1.0, 1.0], device=model.device))) - 1.0) < 1e-5

        test_trainer.fit(model)
        assert math.fabs(float(model(torch.tensor([1.0, 1.0], device=model.device))) - 0.03) < 1e-5

    @pytest.mark.run_only_on('GPU')
    @pytest.mark.parametrize('test_dist_ckpt', [False, True])
    @pytest.mark.pleasefixme
    def test_base_checkpoints_are_not_overwritten(self, tmp_path, test_dist_ckpt):
        """Simulates already existing checkpoints in the ckpt directory and tests non-nemo ckpt versioning"""
        strategy = NLPDDPStrategy() if test_dist_ckpt else 'auto'
        test_trainer = pl.Trainer(
            accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=4, strategy=strategy
        )
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_nemo_on_train_end": True},
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleMCoreModel() if test_dist_ckpt else ExampleModel()

        ckpt_dir = Path(tmp_path / "test" / "checkpoints")
        assert not ckpt_dir.exists()

        # Fake existing 1st and last checkpoint
        suffix = '' if test_dist_ckpt else '.ckpt'
        ckpt_dir.mkdir(parents=True)
        ckpt_1 = ckpt_dir / f'default--val_loss=0.0000-epoch=1{suffix}'
        ckpt_2 = ckpt_dir / f'default--val_loss=0.0300-epoch=2{suffix}'

        if test_dist_ckpt:
            ckpt_1.mkdir()
            with open(ckpt_1 / 'metadata.json', 'w') as f:
                json.dump({'sharded_backend': 'xxx'}, f)
        else:
            ckpt_1.touch()
        # don't create 2nd checkpoint
        ckpt_nemo = ckpt_dir / 'default.nemo'
        ckpt_nemo.touch()

        # Train
        test_trainer.fit(model)

        # Check base checkpoint (without versioning)
        all_checkpoints = [p.name for p in Path(str(tmp_path / "test" / "checkpoints")).glob("*")]
        assert ckpt_1.exists(), all_checkpoints  # existed before
        assert ckpt_2.exists(), all_checkpoints
        assert ckpt_nemo.exists(), all_checkpoints  # existed before

        # Versioned checkpoints
        def _get_versioned_name(ckpt_name: Path, nemo: bool = False):
            if test_dist_ckpt and not nemo:
                # no suffix at all
                return ckpt_name.with_name(ckpt_name.name + '-v1')
            return ckpt_name.with_stem(ckpt_name.stem + '-v1')

        assert _get_versioned_name(ckpt_1).exists(), all_checkpoints
        assert not _get_versioned_name(ckpt_2).exists(), all_checkpoints  # ckpt2 didn't exist before
        # .nemo checkpoints are not versioned:
        assert not _get_versioned_name(ckpt_nemo, nemo=True).exists(), all_checkpoints

    @pytest.mark.unit
    def test_last_checkpoint_saved(self, tmp_path):
        max_steps = 64
        tmp_path = tmp_path / "test_1"

        class TestModel(ExampleModel):
            def train_dataloader(self):
                dataset = OnesDataset(64)
                return torch.utils.data.DataLoader(dataset, batch_size=1)

        trainer = pl.Trainer(
            accelerator='cpu', enable_checkpointing=False, logger=False, max_steps=max_steps, val_check_interval=0.33
        )
        exp_manager(
            trainer,
            {
                "explicit_log_dir": str(tmp_path),
                "checkpoint_callback_params": {"filename": f"{{val_loss:.4f}}-{{epoch}}-{{step}}"},
            },
        )
        model = TestModel()
        trainer.fit(model)

        checkpoint_dir = Path(str(tmp_path / "checkpoints"))
        model_path = checkpoint_dir / "val_loss=0.0300-epoch=1-step=64-last.ckpt"
        last_saved_checkpoint = torch.load(model_path, weights_only=False)
        assert max_steps == last_saved_checkpoint['global_step']

        # restart training, ensure global step starts correctly
        class AssertCallback(Callback):
            def on_train_start(self, trainer: "pl.Trainer", pl_module: "pl.LightningModule") -> None:
                assert trainer.global_step == max_steps

            def on_train_batch_end(
                self, trainer: "pl.Trainer", pl_module: "pl.LightningModule", outputs, batch: Any, batch_idx: int
            ) -> None:
                # we should only be running for one more step.
                assert trainer.global_step == max_steps + 1

        trainer = pl.Trainer(
            accelerator='cpu',
            enable_checkpointing=False,
            logger=False,
            max_steps=65,
            val_check_interval=0.33,
            callbacks=AssertCallback(),
        )
        exp_manager(
            trainer,
            {
                "explicit_log_dir": str(tmp_path),
                "checkpoint_callback_params": {"filename": f"{{val_loss:.4f}}-{{epoch}}-{{step}}"},
            },
        )
        model = TestModel()
        trainer.fit(model, ckpt_path=model_path)

    @pytest.mark.unit
    def test_resume_checkpoint_skip_validation(self, tmp_path):
        """Test to ensure that when we resume from a checkpoint, we do not re-run validation unnecessarily."""
        tmp_path = tmp_path / "test_2"

        def run_training(resume_path=None):
            class TestModel(ExampleModel):
                def train_dataloader(self):
                    dataset = OnesDataset(10)
                    return torch.utils.data.DataLoader(dataset, batch_size=1)

            class AssertCallback(Callback):
                recorded_validations = 0
                recorded_train_steps = 0

                def on_validation_start(self, trainer: "pl.Trainer", pl_module: "pl.LightningModule") -> None:
                    self.recorded_validations += 1

                def on_train_batch_end(
                    self, trainer: "pl.Trainer", pl_module: "pl.LightningModule", outputs, batch: Any, batch_idx: int
                ) -> None:
                    self.recorded_train_steps += 1

                def on_train_end(self, trainer: "pl.Trainer", pl_module: "pl.LightningModule") -> None:
                    if resume_path is not None:
                        # we should only run validation at the end of training.
                        assert self.recorded_validations == 1
                        # we continue from half way
                        assert self.recorded_train_steps == len(pl_module.train_dataloader()) // 2
                    else:
                        # we've run validation within the middle of training and at the end of training.
                        assert self.recorded_validations == 2
                        assert self.recorded_train_steps == len(pl_module.train_dataloader())

            model = TestModel()
            trainer = pl.Trainer(
                accelerator='cpu',
                enable_checkpointing=False,
                logger=False,
                callbacks=[AssertCallback()],
                val_check_interval=0.5,
                num_sanity_val_steps=0,
                max_epochs=1,
            )
            exp_manager(
                trainer,
                {"explicit_log_dir": str(tmp_path), "checkpoint_callback_params": {"filename": f"{{epoch}}-{{step}}"}},
            )
            trainer.fit(model, ckpt_path=resume_path)

        run_training()
        resume_path = tmp_path / 'checkpoints/epoch=0-step=5.ckpt'
        run_training(resume_path)

    def test_warning_validation_skipping_when_custom_epoch_loop(self, tmp_path):
        """When using validation skipping on restart with a custom epoch loop, we warn the user that we skip
        support to not interfere with their custom logic.
        """
        tmp_path = tmp_path / "test_3"

        class CustomLoop(_TrainingEpochLoop): ...

        trainer = pl.Trainer(
            accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1, val_check_interval=0.33
        )
        ## _TrainingEpochLoop in PTL 2.0 takes trainer as an arg
        loop = CustomLoop(trainer)
        trainer.fit_loop.epoch_loop = loop
        with pytest.warns(UserWarning, match="Detected custom epoch loop"):
            exp_manager(trainer, {"explicit_log_dir": str(tmp_path)})

    def _write_fake_checkpoint(self, path, isdir, add_unfinished_marker):
        path = Path(path)
        if isdir:
            # fake distributed checkpoint
            path.mkdir(parents=True, exist_ok=True)
            (path / "dummy.txt").touch()
        else:
            # fake checkpoint file
            path.parent.mkdir(parents=True, exist_ok=True)
            path.touch()
        if add_unfinished_marker:
            NeMoModelCheckpoint.set_checkpoint_unfinished_marker(path)

    @pytest.mark.unit
    def test_skipped_unfinished_checkpoints_when_restoring(self, tmp_path):
        """
        Check if unfinished checkpoints are skipped during last checkpoint lookup.
        Logic of the test:
        - write multiple last checkpoints, some of them incomplete
        - ensure that the last complete checkpoint is found
        """

        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=900-consumed_samples=1000.0.ckpt",
            isdir=False,
            add_unfinished_marker=False,
        )  # not last
        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=900-consumed_samples=1000.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=True,
        )  # incomplete
        self._write_fake_checkpoint(
            checkpoints_dir
            / "mp_rank_00"
            / "megatron_gpt--val_loss=5.01-step=1100-consumed_samples=17600.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=True,
        )  # incomplete
        self._write_fake_checkpoint(
            checkpoints_dir
            / "mp_rank_01"
            / "megatron_gpt--val_loss=5.01-step=1100-consumed_samples=17600.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=True,
        )  # incomplete
        self._write_fake_checkpoint(
            checkpoints_dir
            / "mp_rank_00"
            / "megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=False,
        )  # ok
        self._write_fake_checkpoint(
            checkpoints_dir
            / "mp_rank_01"
            / "megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=False,
        )  # ok

        restored_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        exp_manager(
            restored_trainer,
            {"resume_if_exists": True, "explicit_log_dir": str(test_dir)},
        )

        # Check that last complete (w/o unifinished marker) checkpoint was found
        assert (
            Path(restored_trainer.ckpt_path).name
            == 'megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0-last.ckpt'
        )

    @pytest.mark.unit
    def test_skipped_unfinished_dist_checkpoints_when_restoring(self, tmp_path):
        """
        Check if unfinished distributed checkpoints are skipped during last checkpoint lookup.
        Logic of the test:
        - write multiple last checkpoints, some of them incomplete
        - ensure that the last complete checkpoint is found
        """

        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0",
            isdir=True,
            add_unfinished_marker=False,
        )
        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0-last",
            isdir=True,
            add_unfinished_marker=False,
        )
        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=1100-consumed_samples=17600.0",
            isdir=True,
            add_unfinished_marker=False,
        )
        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=1100-consumed_samples=17600.0-last",
            isdir=True,
            add_unfinished_marker=True,
        )

        restored_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)
        exp_manager(
            restored_trainer,
            {"resume_if_exists": True, "explicit_log_dir": str(test_dir)},
        )

        # Check that last complete (w/o unifinished marker) checkpoint was found
        assert (
            Path(restored_trainer.ckpt_path).name
            == 'megatron_gpt--val_loss=5.01-step=1000-consumed_samples=16000.0-last'
        )

    @pytest.mark.unit
    def test_incomplete_checkpoints_cleanup(self, tmp_path):
        """
        Check if unfinished checkpoints are cleaned up when training starts
        Complete checkpoints should be left intact.
        """
        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        complete_ckpts = {
            checkpoints_dir / "step=1-epoch=0.ckpt",
            checkpoints_dir / "step=2-epoch=0-last.ckpt",
            checkpoints_dir / "mp_rank_00" / "step=3-epoch=0-last.ckpt",
            checkpoints_dir / "tp_rank_00_pp_rank_000" / "step=4-epoch=0-last.ckpt",
            checkpoints_dir / "tp_rank_00_pp_rank_001" / "step=4-epoch=0-last.ckpt",
        }
        for ckpt_filepath in complete_ckpts:
            self._write_fake_checkpoint(ckpt_filepath, isdir=False, add_unfinished_marker=False)

        incomplete_ckpts = {
            checkpoints_dir / "step=11-epoch=1.ckpt",
            checkpoints_dir / "step=12-epoch=1-last.ckpt",
            checkpoints_dir / "mp_rank_00" / "step=13-epoch=1-last.ckpt",
            checkpoints_dir / "tp_rank_00_pp_rank_000" / "step=14-epoch=1-last.ckpt",
            checkpoints_dir / "tp_rank_00_pp_rank_001" / "step=14-epoch=1-last.ckpt",
        }
        for ckpt_filepath in incomplete_ckpts:
            self._write_fake_checkpoint(ckpt_filepath, isdir=False, add_unfinished_marker=True)

        # sanity check
        remaining_ckpts = {f for f in (test_dir / "checkpoints").rglob("*.ckpt") if f.is_file()}
        assert remaining_ckpts == (complete_ckpts | incomplete_ckpts)

        # marker without corresponding checkpoint should be removed during cleanup in exp_manager
        (checkpoints_dir / f"orphan-marker001-{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}").touch()

        # unfinished checkpoint with EMA part, both parts should be removed
        self._write_fake_checkpoint(
            checkpoints_dir / "incomplete01-EMA.ckpt",
            isdir=False,
            add_unfinished_marker=False,
        )
        self._write_fake_checkpoint(checkpoints_dir / "incomplete01.ckpt", isdir=False, add_unfinished_marker=True)

        # just EMA part - should be removed. NOTE marker path is the same for base part and for EMA part
        self._write_fake_checkpoint(
            checkpoints_dir / "incomplete02-EMA.ckpt",
            isdir=False,
            add_unfinished_marker=False,
        )
        (checkpoints_dir / f"incomplete02{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}").touch()

        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)

        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_top_k": 0, "save_last": False},
                "explicit_log_dir": str(test_dir),
            },
        )

        model = ExampleModel()
        test_trainer.fit(model)

        remaining_ckpts = {f for f in (test_dir / "checkpoints").rglob("*.ckpt") if f.is_file()}
        assert remaining_ckpts == complete_ckpts
        remaining_markers = list(checkpoints_dir.rglob(f"*{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}"))
        assert remaining_markers == []

    @pytest.mark.unit
    def test_incomplete_dist_checkpoints_cleanup(self, tmp_path):
        """
        Check if unfinished distributed checkpoints are cleaned up when training starts.
        Complete distributed checkpoints should be left intact.
        """

        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        complete_dist_ckpts = {
            checkpoints_dir / "step=5-epoch=0",
            checkpoints_dir / "step=6-epoch=0-last",
        }
        for ckpt_dirpath in complete_dist_ckpts:
            self._write_fake_checkpoint(ckpt_dirpath, isdir=True, add_unfinished_marker=False)

        incomplete_dist_ckpts = {
            checkpoints_dir / "step=15-epoch=1",
            checkpoints_dir / "step=16-epoch=1-last",
        }
        for ckpt_dirpath in incomplete_dist_ckpts:
            self._write_fake_checkpoint(ckpt_dirpath, isdir=True, add_unfinished_marker=True)

        # marker without corresponding checkpoint should be removed during cleanup in exp_manager
        (checkpoints_dir / f"orphan-marker001-{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}").touch()

        remaining_dist_ckpts = {f for f in (test_dir / "checkpoints").glob("*") if f.is_dir()}
        assert remaining_dist_ckpts == (complete_dist_ckpts | incomplete_dist_ckpts)

        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)

        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {"save_top_k": 0, "save_last": False},
                "explicit_log_dir": str(test_dir),
            },
        )

        model = ExampleModel()
        test_trainer.fit(model)

        remaining_dist_ckpts = {f for f in (test_dir / "checkpoints").glob("*") if f.is_dir()}
        assert remaining_dist_ckpts == complete_dist_ckpts
        remaining_markers = list(checkpoints_dir.rglob(f"*{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}"))
        assert remaining_markers == []

    _chkpt_path_and_marker_path_pairs = [
        ('a=1_b=1.c.d.e', f'a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
        ('a=1_b=1.c.d.e-last', f'a=1_b=1.c.d.e-last{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
        ('.ckpt/a=1_b=1.c.d.e.ckpt', f'.ckpt/a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
        ('.ckpt/a=1_b=1.c.d.e-EMA.ckpt', f'.ckpt/a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
        (
            '.ckpt/a=1_b=1.c.d.e-last.ckpt',
            f'.ckpt/a=1_b=1.c.d.e-last{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}',
        ),
        (
            '/tmp/mp_rank_00/a=1_b=1.c.d.e.ckpt',
            f'/tmp/a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}',
        ),
        (
            '/tmp/tp_rank_00_pp_rank_000/a=1_b=1.c.d.e.ckpt',
            f'/tmp/a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}',
        ),
        ('nemo/a=1_b=1.c.d.e.nemo', f'nemo/a=1_b=1.c.d.e{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
        ('nemo/a=1_b=1.c.d.e-last.nemo', f'nemo/a=1_b=1.c.d.e-last{NeMoModelCheckpoint.UNFINISHED_CHECKPOINT_SUFFIX}'),
    ]

    @pytest.mark.unit
    @pytest.mark.parametrize("chkpt_path, expected_marker_path", _chkpt_path_and_marker_path_pairs)
    def test_incomplete_checkpoints_marker_path(self, chkpt_path, expected_marker_path):
        """
        Ensure that unfinished checkpoint marker path is correctly formed.
        """
        marker_path = NeMoModelCheckpoint.format_checkpoint_unfinished_marker_path(chkpt_path)
        assert str(marker_path) == str(expected_marker_path)

    @pytest.mark.unit
    def test_invalid_checkpoints_removed_from_topk(self, tmp_path):
        """
        Ensure that invalid (unfinished, deleted) checkpoints are removed from topk when resuming.
        - Do few training steps and save checkpoints
        - Delete some checkpoints, mark some as unfinished
        - Resume training and verify that topk checkpoints are correct
        """
        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=7)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {
                    "save_top_k": 3,
                    "save_last": True,
                    "mode": 'max',
                    "monitor": 'epoch',
                    "filename": f"{{epoch}}",
                },
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer.fit(model)

        ckpt_filenames = {f.name for f in checkpoints_dir.rglob("*.ckpt") if f.is_file()}
        assert len(ckpt_filenames) == 4  # 3 top + 1 last
        assert 'epoch=7-last.ckpt' in ckpt_filenames
        assert 'epoch=6.ckpt' in ckpt_filenames
        assert 'epoch=5.ckpt' in ckpt_filenames
        assert 'epoch=4.ckpt' in ckpt_filenames

        # Mark 6th epoch checkpoint as unfinished and remove 5th epoch checkpoint,
        # so last valid candidate for topk is 4th epoch checkpoint
        NeMoModelCheckpoint.set_checkpoint_unfinished_marker(checkpoints_dir / 'epoch=6.ckpt')
        (checkpoints_dir / 'epoch=5.ckpt').unlink()

        test_trainer2 = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=9)
        exp_manager(
            test_trainer2,
            {
                "resume_if_exists": True,
                "checkpoint_callback_params": {
                    "save_top_k": 3,
                    "save_last": True,
                    "mode": 'max',
                    "monitor": 'epoch',
                    "filename": f"{{epoch}}",
                },
                "explicit_log_dir": str(tmp_path / "test"),
            },
        )
        model = ExampleModel()
        test_trainer2.fit(model)

        ckpt_filenames = {f.name for f in checkpoints_dir.rglob("*.ckpt") if f.is_file()}
        # 3 top + 1 last
        assert len(ckpt_filenames) == 4
        assert 'epoch=9-last.ckpt' in ckpt_filenames
        assert 'epoch=8.ckpt' in ckpt_filenames
        assert 'epoch=7.ckpt' in ckpt_filenames
        assert 'epoch=4.ckpt' in ckpt_filenames

    @pytest.mark.unit
    def test_doesnt_silently_start_from_scratch(self, tmp_path):
        """
        Ensure that if the last checkpoint is unfinished it wont silently start from scratch.
        This is to avoid a training that is not actually making any progress.
        """
        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=900-consumed_samples=1000.0-last.ckpt",
            isdir=False,
            add_unfinished_marker=True,
        )  # incomplete last

        restored_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)

        with pytest.raises(Exception):
            exp_manager(
                restored_trainer,
                {"resume_if_exists": True, "resume_ignore_no_checkpoint": True, "explicit_log_dir": str(test_dir)},
            )

    @pytest.mark.unit
    def test_doesnt_silently_start_from_scratch_dist(self, tmp_path):
        """
        Ensure that if the last distributed checkpoint is unfinished it wont silently start from scratch.
        This is to avoid a training that is not actually making any progress.
        """

        test_dir = tmp_path / "test"
        checkpoints_dir = test_dir / "checkpoints"

        self._write_fake_checkpoint(
            checkpoints_dir / "megatron_gpt--val_loss=5.01-step=1100-consumed_samples=17600.0-last",
            isdir=True,
            add_unfinished_marker=True,
        )  # incomplete last

        restored_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False)

        with pytest.raises(Exception):
            exp_manager(
                restored_trainer,
                {"resume_if_exists": True, "resume_ignore_no_checkpoint": True, "explicit_log_dir": str(test_dir)},
            )

    @pytest.mark.unit
    def test_save_nemo_not_comp_with_model_parallel(self, tmp_path):
        """
        Ensure that always_save_nemo is not compatible with model parallelism.
        """

        test_dir = tmp_path / "test"

        with pytest.raises(LoggerMisconfigurationError):
            appstate = AppState()
            appstate.tensor_model_parallel_size = 2
            appstate.pipeline_model_parallel_size = 1
            appstate.context_parallel_size = 1
            test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)
            exp_manager(
                test_trainer,
                {
                    "checkpoint_callback_params": {
                        "always_save_nemo": True,
                    },
                    "explicit_log_dir": str(test_dir),
                },
            )

        with pytest.raises(LoggerMisconfigurationError):
            appstate = AppState()
            appstate.tensor_model_parallel_size = 1
            appstate.pipeline_model_parallel_size = 2
            appstate.context_parallel_size = 1
            test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)
            exp_manager(
                test_trainer,
                {
                    "checkpoint_callback_params": {
                        "always_save_nemo": True,
                    },
                    "explicit_log_dir": str(test_dir),
                },
            )

        with pytest.raises(LoggerMisconfigurationError):
            appstate = AppState()
            appstate.tensor_model_parallel_size = 1
            appstate.pipeline_model_parallel_size = 1
            appstate.context_parallel_size = 2
            test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)
            exp_manager(
                test_trainer,
                {
                    "checkpoint_callback_params": {
                        "always_save_nemo": True,
                    },
                    "explicit_log_dir": str(test_dir),
                },
            )

        appstate = AppState()
        appstate.tensor_model_parallesl_size = 1
        appstate.pipeline_model_parallel_size = 1
        appstate.context_parallel_size = 1
        test_trainer = pl.Trainer(accelerator='cpu', enable_checkpointing=False, logger=False, max_epochs=1)
        exp_manager(
            test_trainer,
            {
                "checkpoint_callback_params": {
                    "always_save_nemo": True,
                },
                "explicit_log_dir": str(test_dir),
            },
        )