deviandice/DL-Art-School
1
0
Fork 0
forked from mrq/DL-Art-School
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Load ema to cpu memory if specified

Browse Source
This commit is contained in:
James Betker 2022-01-24 15:08:29 -07:00
parent 49edffb6ad
commit dfef34ba39
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
codes/trainer/ExtensibleTrainer.py
View File

@ -45,6 +45,9 @@ class ExtensibleTrainer(BaseModel):
self.env['mega_batch_factor'] = self.mega_batch_factor
self.batch_factor = self.mega_batch_factor
self.ema_rate = opt_get(train_opt, ['ema_rate'], .999)
# It is advantageous for large networks to do this to save an extra copy of the model weights.
# It does come at the cost of a round trip to CPU memory at every batch.
self.ema_on_cpu = opt_get(train_opt, ['ema_on_cpu'], False)
self.checkpointing_cache = opt['checkpointing_enabled']
self.auto_recover = opt_get(opt, ['automatically_recover_nan_by_reverting_n_saves'], None)
@ -147,6 +150,8 @@ class ExtensibleTrainer(BaseModel):
self.networks[k] = dnet
if self.is_train:
self.emas[k] = copy.deepcopy(v)
if self.ema_on_cpu:
self.emas[k] = self.emas[k].cpu()
found += 1
assert found == len(self.netsG) + len(self.netsD)
@ -316,6 +321,8 @@ class ExtensibleTrainer(BaseModel):
ema_params = self.emas[name].parameters()
net_params = net.parameters()
for ep, np in zip(ema_params, net_params):
if self.ema_on_cpu:
np = np.cpu()
ep.detach().mul_(self.ema_rate).add_(np, alpha=1 - self.ema_rate)
[e.after_optimize(state) for e in self.experiments]
@ -443,6 +450,8 @@ class ExtensibleTrainer(BaseModel):
else:
print("WARNING! Unable to find EMA network! Starting a new EMA from given model parameters.")
self.emas[name] = copy.deepcopy(net)
if self.ema_on_cpu:
self.emas[name] = self.emas[name].cpu()
if hasattr(net.module, 'network_loaded'):
net.module.network_loaded()