wassname/pytorch-lightning
1
0
Fork 0
mirror of https://github.com/wassname/pytorch-lightning.git synced 2026-09-12 12:40:20 +08:00
Code Issues Packages Projects Releases Wiki Activity

added single node distdataparallel

Browse Source
This commit is contained in:
William Falcon
2019-07-03 16:21:56 -04:00
parent 62774ffacb
commit 8d3090c843
1 changed files with 4 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pytorch_lightning/models/trainer.py
+4 -1
View File
@@ -281,7 +281,7 @@ class Trainer(TrainerIO):
if self.on_gpu:
rank = 0
self.experiment = self.experiment.get_meta_copy()
mp.spawn(dummy, nprocs=len(self.data_parallel_device_ids), args=(self, ))
mp.spawn(self.__dp_train, nprocs=len(self.data_parallel_device_ids), args=(self, ))
else:
self.__run_pretrain_routine(model)
@@ -299,11 +299,14 @@ class Trainer(TrainerIO):
:param cluster_obj:
:return:
"""
print('in process')
print(self.experiment)
# TODO: pass in ip
ip = "127.0.0.1"
print(self.data_parallel_device_ids)
# configure server
rank = proc_rank * len(self.data_parallel_device_ids) + gpu_nb
print(f"GPU: {gpu_nb} - Rank: {rank}")