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

PERF: vectorise for loop using torch-native functions #137

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

PERF: vectorise for loop using torch-native functions #137

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
b2c5dc0
PERF: vectorise for loop using torch-native functions
sbrugman Sep 9, 2024
dbe5284
Correct function naming from benchmark
sbrugman Sep 14, 2024
c41dcaa
Re-add comment lost in benchmark
sbrugman Sep 14, 2024
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
27 changes: 12 additions & 15 deletions models/llama3/reference_impl/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,7 +42,7 @@ def forward(self, x):
return output * self.weight


def apply_scaling(freqs: torch.Tensor):
def apply_scaling(freqs: torch.Tensor) -> torch.Tensor:
# Values obtained from grid search
scale_factor = 8
low_freq_factor = 1
Expand All @@ -51,20 +51,17 @@ def apply_scaling(freqs: torch.Tensor):

low_freq_wavelen = old_context_len / low_freq_factor
high_freq_wavelen = old_context_len / high_freq_factor
new_freqs = []
for freq in freqs:
wavelen = 2 * math.pi / freq
if wavelen < high_freq_wavelen:
new_freqs.append(freq)
elif wavelen > low_freq_wavelen:
new_freqs.append(freq / scale_factor)
else:
assert low_freq_wavelen != high_freq_wavelen
smooth = (old_context_len / wavelen - low_freq_factor) / (
high_freq_factor - low_freq_factor
)
new_freqs.append((1 - smooth) * freq / scale_factor + smooth * freq)
return torch.tensor(new_freqs, dtype=freqs.dtype, device=freqs.device)

wavelen = 2 * torch.pi / freqs
new_freqs = torch.where(wavelen > low_freq_wavelen, freqs / scale_factor, freqs)
smooth = (old_context_len / wavelen - low_freq_factor) / (
high_freq_factor - low_freq_factor
)
return torch.where(
(wavelen >= high_freq_wavelen) & (wavelen <= low_freq_wavelen),
(1 - smooth) * new_freqs / scale_factor + smooth * new_freqs,
new_freqs,
)


def precompute_freqs_cis(
Expand Down