Skip to content

Support bias for generic_gemm #70

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

Draft
wenqingqian wants to merge 2 commits into
base: main
Choose a base branch
from
+12 −3
Draft

Support bias for generic_gemm #70

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2ab41b8
bias linear
wenqingqian May 27, 2026
4677c76
minor
wenqingqian May 27, 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
15 changes: 12 additions & 3 deletions transformer_engine/plugin/core/backends/flagos/impl/gemm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,7 +71,6 @@ def generic_gemm_fl(

assert not gelu and gelu_in is None, "Triton-Based General Gemm do not support gelu now"
assert quantizer is None, "Triton-Based General Gemm do not support quantization now"
assert bias is None, "Triton-Based General Gemm do not support bias now"

alpha = validate_gemm_scale(alpha, True)
beta = validate_gemm_scale(beta, accumulate)
Expand All @@ -95,7 +94,18 @@ def generic_gemm_fl(
A_comp = A.T if transA else A
B_comp = B.T if transB else B

out1 = flag_gems.mm(B_comp, A_comp)
bias_grad = None
if grad:
out1 = flag_gems.mm(B_comp, A_comp)
if bias is not None:
bias_grad = flag_gems.sum_dim(B, dim=[0])
else:
# NOTE(wqq) flag_gems.addmm uses beta for bias scaling (Y = alpha * WX + beta * bias),
# unlike the beta here (for scaling D). Always set to 1.
if bias is not None:
out1 = flag_gems.addmm(bias, B_comp, A_comp, beta=1, alpha=alpha)
else:
out1 = flag_gems.mm(B_comp, A_comp)

if shape_b_changed:
out1 = out1.view(s, b, -1)
Expand All @@ -104,7 +114,6 @@ def generic_gemm_fl(
if torch_out_dtype is not None and out1.dtype != torch_out_dtype:
out1 = out1.to(torch_out_dtype)

bias_grad = None
gelu_input = None
extra_output_ret = None

Expand Down
Loading