Skip to content

[CI] fix bench_jit_per_token_group_quant_8bit sglang import cascade #453

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

[CI] fix bench_jit_per_token_group_quant_8bit sglang import cascade #453

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
fd910ae
[CI] re-enable bench_jit_per_token_group_quant_8bit; pin enable_v2=False
arathi-hlab Aug 27, 2026
9dc42e0
retrigger CI
arathi-hlab Aug 27, 2026
32aa6b5
Merge branch 'main' into ci/fix-jit-8bit-bench-sglang-import
arathi-hlab Aug 28, 2026
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: 7 additions & 2 deletions benchmark/bench_jit_per_token_group_quant_8bit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,10 +63,15 @@ def aot_per_token_group_quant_8bit(x, group_size, dst_dtype, eps=1e-10):
x_q = torch.empty_like(x, dtype=dst_dtype)
x_s = torch.empty((m, n // group_size), device=x.device, dtype=torch.float32)
min_8bit, max_8bit = _minmax(dst_dtype)
# Pin v1; enable_v2=None would import sglang (not in bench container).
if dst_dtype == torch.int8:
sgl_per_token_group_quant_int8(x, x_q, x_s, group_size, eps, min_8bit, max_8bit)
sgl_per_token_group_quant_int8(
x, x_q, x_s, group_size, eps, min_8bit, max_8bit, enable_v2=False
)
else:
sgl_per_token_group_quant_fp8(x, x_q, x_s, group_size, eps, min_8bit, max_8bit)
sgl_per_token_group_quant_fp8(
x, x_q, x_s, group_size, eps, min_8bit, max_8bit, enable_v2=False
)
return x_q, x_s


Expand Down
Loading