def test_cpu_slurm_save_load():
"""
Verify model save/load/checkpoint on CPU
:return:
"""
hparams = get_hparams()
model = LightningTestModel(hparams)
save_dir = init_save_dir()
# exp file to get meta
exp = get_exp(False)
exp.argparse(hparams)
exp.save()
version = exp.version
trainer_options = dict(max_nb_epochs=1,
experiment=exp,
checkpoint_callback=ModelCheckpoint(save_dir))
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
real_global_step = trainer.global_step
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# predict with trained model before saving
# make a prediction
for batch in model.test_dataloader:
break
x, y = batch
x = x.view(x.size(0), -1)
model.eval()
pred_before_saving = model(x)
# test HPC saving
# simulate snapshot on slurm
saved_filepath = trainer.hpc_save(save_dir, exp)
assert os.path.exists(saved_filepath)
# new exp file to get meta
exp = get_exp(False, version=version)
exp.argparse(hparams)
exp.save()
trainer_options = dict(
max_nb_epochs=1,
experiment=exp,
checkpoint_callback=ModelCheckpoint(save_dir),
)
trainer = Trainer(**trainer_options)
model = LightningTestModel(hparams)
# set the epoch start hook so we can predict before the model does the full training
def assert_pred_same():
assert trainer.global_step == real_global_step and trainer.global_step > 0
# predict with loaded model to make sure answers are the same
trainer.model.eval()
new_pred = trainer.model(x)
assert torch.all(torch.eq(pred_before_saving, new_pred)).item() == 1
model.on_epoch_start = assert_pred_same
# by calling fit again, we trigger training, loading weights from the cluster
# and our hook to predict using current model before any more weight updates
trainer.fit(model)
clear_save_dir()
def test_cpu_slurm_save_load():
"""
Verify model save/load/checkpoint on CPU
:return:
"""
hparams = get_hparams()
model = LightningTestModel(hparams)
save_dir = init_save_dir()
# exp file to get meta
exp = get_exp(False)
exp.argparse(hparams)
exp.save()
cluster_a = SlurmCluster()
trainer_options = dict(
max_nb_epochs=1,
cluster=cluster_a,
experiment=exp,
checkpoint_callback=ModelCheckpoint(save_dir)
)
# fit model
trainer = Trainer(**trainer_options)
result = trainer.fit(model)
real_global_step = trainer.global_step
# traning complete
assert result == 1, 'amp + ddp model failed to complete'
# predict with trained model before saving
# make a prediction
for batch in model.test_dataloader:
break
x, y = batch
x = x.view(x.size(0), -1)
model.eval()
pred_before_saving = model(x)
# test registering a save function
trainer.enable_auto_hpc_walltime_manager()
# test HPC saving
# simulate snapshot on slurm
saved_filepath = trainer.hpc_save(save_dir, exp)
assert os.path.exists(saved_filepath)
# wipe-out trainer and model
# retrain with not much data... this simulates picking training back up after slurm
# we want to see if the weights come back correctly
continue_tng_hparams = get_hparams(continue_training=True,
hpc_exp_number=cluster_a.hpc_exp_number)
trainer_options = dict(
max_nb_epochs=1,
cluster=SlurmCluster(continue_tng_hparams),
experiment=exp,
checkpoint_callback=ModelCheckpoint(save_dir),
)
trainer = Trainer(**trainer_options)
model = LightningTestModel(hparams)
# set the epoch start hook so we can predict before the model does the full training
def assert_pred_same():
assert trainer.global_step == real_global_step and trainer.global_step > 0
# predict with loaded model to make sure answers are the same
trainer.model.eval()
new_pred = trainer.model(x)
assert torch.all(torch.eq(pred_before_saving, new_pred)).item() == 1
model.on_epoch_start = assert_pred_same
# by calling fit again, we trigger training, loading weights from the cluster
# and our hook to predict using current model before any more weight updates
trainer.fit(model)
clear_save_dir()
