Rooba/lightning
1
0
Fork 0
mirror of https://github.com/Lightning-AI/lightning.git synced 2025-02-26 20:25:16 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update single_gpu_node_ddp_template.py

Browse Source
This commit is contained in:
William Falcon 2019-10-05 13:55:05 -04:00 committed by GitHub
parent 0e2b0e39b5
commit c86524b0cc
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
examples/new_project_templates/single_gpu_node_ddp_template.py
View File

@ -25,36 +25,18 @@ def main(hparams):
# ------------------------ # ------------------------
# 1 INIT LIGHTNING MODEL # 1 INIT LIGHTNING MODEL
# ------------------------ # ------------------------
print('loading model...')
model = LightningTemplateModel(hparams) model = LightningTemplateModel(hparams)
print('model built')
# ------------------------ # ------------------------
# 2 INIT TEST TUBE EXP # 2 INIT TRAINER
# ------------------------
# init experiment
exp = Experiment(
name=hyperparams.experiment_name,
save_dir=hyperparams.test_tube_save_path,
autosave=False,
description='test demo'
)
exp.argparse(hparams)
exp.save()
# ------------------------
# 3 INIT TRAINER
# ------------------------ # ------------------------
trainer = Trainer( trainer = Trainer(
experiment=exp,
gpus=hparams.gpus, gpus=hparams.gpus,
distributed_backend=hparams.dist_backend distributed_backend=hparams.dist_backend
) )
# ------------------------ # ------------------------
# 4 START TRAINING # 3 START TRAINING
# ------------------------ # ------------------------
trainer.fit(model) trainer.fit(model)
@ -76,12 +58,6 @@ if __name__ == '__main__':
' value -1 uses all the gpus on the node') ' value -1 uses all the gpus on the node')
parent_parser.add_argument('--dist_backend', type=str, default='ddp', parent_parser.add_argument('--dist_backend', type=str, default='ddp',
help='When using multiple GPUs set Trainer(distributed_backend=dp) (or ddp)') help='When using multiple GPUs set Trainer(distributed_backend=dp) (or ddp)')
parent_parser.add_argument('--test_tube_save_path', type=str, default=test_tube_dir,
help='where to save logs')
parent_parser.add_argument('--model_save_path', type=str, default=checkpoint_dir,
help='where to save model')
parent_parser.add_argument('--experiment_name', type=str, default='pt_lightning_exp_a',
help='test tube exp name')
# allow model to overwrite or extend args # allow model to overwrite or extend args
parser = LightningTemplateModel.add_model_specific_args(parent_parser, root_dir) parser = LightningTemplateModel.add_model_specific_args(parent_parser, root_dir)
@ -90,6 +66,4 @@ if __name__ == '__main__':
# --------------------- # ---------------------
# RUN TRAINING # RUN TRAINING
# --------------------- # ---------------------
# run on HPC cluster
print(f'RUNNING INTERACTIVE MODE ON GPUS. gpu ids: {hyperparams.gpus}')
main(hyperparams) main(hyperparams)