Skip to content

Added type fix in the docstrings of PagedAttention #875

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

Open
AdityaKulshrestha wants to merge 2 commits into
base: main
Choose a base branch
from
+8 −8
Open

Added type fix in the docstrings of PagedAttention #875

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
353d7d1
Added typo fix
AdityaKulshrestha Dec 21, 2025
68bc3b8
linting fixes
AdityaKulshrestha Dec 21, 2025
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
2 changes: 1 addition & 1 deletion examples/cpu/features/int8_recipe_tuning/int8_autotune.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,7 +84,7 @@ def train(dataloader, model, loss_fn, optimizer):

epochs = 5
for t in range(epochs):
print(f"Epoch {t+1}\n-------------------------------")
print(f"Epoch {t + 1}\n-------------------------------")
train(train_dataloader, model, loss_fn, optimizer)
print("Done!")

Expand Down
4 changes: 2 additions & 2 deletions examples/cpu/inference/python/models/bert_large/training/cpu/run_pretrain_mlperf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1158,8 +1158,8 @@ def main():
print(
f"Step {training_steps:5d}: loss: {gloss:6.3f} lm_acc: {lm_acc:.3f} \
seq_acc: {seq_acc:.3f} lbs: {args.train_batch_size} gbs: {total_batch_size} \
DT: {(t1-t0)*1000.0:.1f} XT: {(t2-t1)*1000.0:.1f} FT: {(t3-t2)*1000.0:.1f} \
BT: {(t4-t3)*1000.0:.1f} OT: {(t5-t4)*1000.0:.1f} TT: {(t5-t0)*1000.0:.1f}"
DT: {(t1 - t0) * 1000.0:.1f} XT: {(t2 - t1) * 1000.0:.1f} FT: {(t3 - t2) * 1000.0:.1f} \
BT: {(t4 - t3) * 1000.0:.1f} OT: {(t5 - t4) * 1000.0:.1f} TT: {(t5 - t0) * 1000.0:.1f}"
)

update_step = training_steps % args.gradient_accumulation_steps == 0
Expand Down
2 changes: 1 addition & 1 deletion examples/cpu/llm/inference/distributed/run_generation_with_deepspeed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1064,7 +1064,7 @@ def trace_handler(prof):
generated, _ = generate()
t_generate_span = time.time() - t_generate_start
for i, o, _ in generated:
print_rank0(f"{'-'*60}\nin={i}\nout={o}\n")
print_rank0(f"{'-' * 60}\nin={i}\nout={o}\n")

# benchmark it!
else:
Expand Down
8 changes: 4 additions & 4 deletions intel_extension_for_pytorch/llm/modules/mha_fusion.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -553,12 +553,12 @@ class PagedAttention:
alibi_slopes (torch.Tensor, optinal): which is the alibi slope with the shape of (num_heads).
softcap (float): the positive softcap value to apply on the attention weights, default is -1.
[class method]: flash_atten_varlen
[class method]: flash_attn_varlen_func
.. highlight:: python
.. code-block:: python
ipex.llm.modules.PagedAttention.flash_atten_varlen(
ipex.llm.modules.PagedAttention.flash_attn_varlen_func(
out,
query,
key_cache,
Expand All @@ -573,8 +573,8 @@ class PagedAttention:
alibi_slopes,
window_size_left,
window_size_right,
k_scale,
v_scale
k_scale=k_scale,
v_scale=v_scale
)
Args:
Expand Down