Skip to content

update to add setting for fixing a seed before execution of f and g f… #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
RobinBruegger merged 2 commits into from
Jan 23, 2020
Merged

update to add setting for fixing a seed before execution of f and g f… #4

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0cb733b
update to add setting for fixing a seed before execution of f and g f…
lucidrains Jan 23, 2020
89cb849
add fix_random_seed keyword attribute when instantiating ReversibleBlock
lucidrains Jan 23, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 16 additions & 1 deletion revtorch/revtorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,8 @@
import torch
import torch.nn as nn
#import torch.autograd.function as func
import sys
import random

class ReversibleBlock(nn.Module):
'''
Expand All @@ -15,11 +17,20 @@ class ReversibleBlock(nn.Module):
g_block (nn.Module): arbitrary subnetwork whos output shape is equal to its input shape
'''

def __init__(self, f_block, g_block, split_along_dim=1):
def __init__(self, f_block, g_block, split_along_dim=1, fix_random_seed = False):
super(ReversibleBlock, self).__init__()
self.f_block = f_block
self.g_block = g_block
self.split_along_dim = split_along_dim
self.fix_random_seed = fix_random_seed
self.random_seeds = {}

def set_seed(self, namespace, new = False):
if not self.fix_random_seed:
return
if new:
self.random_seeds[namespace] = random.randint(0, sys.maxsize)
torch.manual_seed(self.random_seeds[namespace])

def forward(self, x):
"""
Expand All @@ -30,7 +41,9 @@ def forward(self, x):
x1, x2 = torch.chunk(x, 2, dim=self.split_along_dim)
y1, y2 = None, None
with torch.no_grad():
self.set_seed('f', new=True)
y1 = x1 + self.f_block(x2)
self.set_seed('g', new=True)
y2 = x2 + self.g_block(y1)

return torch.cat([y1, y2], dim=self.split_along_dim)
Expand Down Expand Up @@ -64,6 +77,7 @@ def backward_pass(self, y, dy):

# Ensures that PyTorch tracks the operations in a DAG
with torch.enable_grad():
self.set_seed('g')
gy1 = self.g_block(y1)

# Use autograd framework to differentiate the calculation. The
Expand All @@ -83,6 +97,7 @@ def backward_pass(self, y, dy):

with torch.enable_grad():
x2.requires_grad = True
self.set_seed('f')
fx2 = self.f_block(x2)

# Use autograd framework to differentiate the calculation. The
Expand Down