Skip to content
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

Jump to bottom

Fix caching bug in causal convolutions for cache-aware ASR models #7034

Merged
merged 8 commits into from
Jul 20, 2023
Merged

Fix caching bug in causal convolutions for cache-aware ASR models #7034

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
c7df539
fixed the caching of convolutions.
VahidooX Jul 14, 2023
abcc354
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jul 14, 2023
fc25e11
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
VahidooX Jul 14, 2023
8878bf4
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
VahidooX Jul 14, 2023
0d95178
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
VahidooX Jul 18, 2023
c78ef4d
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
borisfom Jul 18, 2023
74b0ae3
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
VahidooX Jul 18, 2023
1df407d
Merge branch 'r1.20.0' into fix_cacheaware_conv_cache
VahidooX Jul 20, 2023
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
9 changes: 6 additions & 3 deletions nemo/collections/asr/parts/submodules/causal_convs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -130,13 +130,16 @@ def __init__(
def update_cache(self, x, cache=None):
if cache is None:
new_x = F.pad(x, pad=(self._left_padding, self._right_padding))
next_cache = cache
else:
new_x = F.pad(x, pad=(0, self._right_padding))
new_x = torch.cat([cache, new_x], dim=-1)
if self.cache_drop_size > 0:
x = x[:, :, : -self.cache_drop_size]
cache = torch.cat([cache[:, :, x.size(-1) :], x], dim=-1)
return new_x, cache
next_cache = new_x[:, :, : -self.cache_drop_size]
else:
next_cache = new_x
next_cache = next_cache[:, :, -cache.size(-1) :]
return new_x, next_cache

def forward(self, x, cache=None):
x, cache = self.update_cache(x, cache=cache)
Expand Down