Morph MORPH
SIGN IN SIGN UP
dhruvhead / ao UNCLAIMED
0 0 1 Python

COMMITS

/ benchmarks/float8/float8_roofline.py
docs
August 21, 2025
V
add simple roofline for float inference with rowwise scaling (#2819)
Vasiliy Kuznetsov committed
b6435f9
August 12, 2025
V
fix float8 training benchmarks on AMD (#2737)
Vasiliy Kuznetsov committed
c88ebe8
August 4, 2025
V
mx: expose scaling calculation methods in training UX (#2620)
Vasiliy Kuznetsov committed
c993d64
July 22, 2025
V
minor cleanup to float8 roofline script (#2579)
Vasiliy Kuznetsov committed
74808e2
June 24, 2025
V
rename `torchao.testing.float8` to `torchao.testing.training` (#2415)
Vasiliy Kuznetsov committed
1a701e6
April 22, 2025
J
Update ruff version to 0.11.6 (#2103)
Jerry Zhang committed
5b775fe
March 28, 2025
V
mx_formats: move training to the quantize_ API (#1970)
Vasiliy Kuznetsov committed
ce759d5
March 18, 2025
M
Add MI300X specs to roofline benchmark (#1913)
Matthias Reso committed
2db0292
March 11, 2025
V
enable mxfp8_cublas recipe in roofline script (#1843)
Vasiliy Kuznetsov committed
d3b481b
March 4, 2025
V
roofline estimator: add float8 rowwise and mxfp8 recipe support (#1789)
Vasiliy Kuznetsov committed
d8af7d7
February 27, 2025
V
roofline estimator: simplify (#1783)
Vasiliy Kuznetsov committed
f478692
V
roofline estimation: delete axiswise scaling, for now (#1782)
Vasiliy Kuznetsov committed
cd69415
V
roofline estimation: delete scaling type (#1781)
Vasiliy Kuznetsov committed
e6706ca
V
[1/x] mx roofline: make the script work on NVIDIA B200 (#1778)
Vasiliy Kuznetsov committed
c788ee7
February 26, 2025
V
[2/x] float8 cleanup: move roofline utils to testing (#1780)
Vasiliy Kuznetsov committed
d00ee41
February 22, 2025
V
delete delayed scaling from torchao.float8 (#1753)
Vasiliy Kuznetsov committed
d370196
February 19, 2025
V
float8 training: make the "config from recipe" API polished (#1731)
Vasiliy Kuznetsov committed
c6c388b
V
float8 training: clean up recipe names (#1730)
Vasiliy Kuznetsov committed
7fc8ad4
January 8, 2025
A
Lint benchmark folder (#1519)
Apurva Jain committed
e4827f2
October 7, 2024
V
float8 training axiswise scaling support with per-gemm-argument configuration (#940)
Vasiliy Kuznetsov committed
dec0313
September 4, 2024
V
make f8 roofline script calculate observed overhead (#734)
Vasiliy Kuznetsov committed
9d92434
August 22, 2024
V
f8 roofline: make utils reusable (#731)
Vasiliy Kuznetsov committed
b09307a
August 15, 2024
V
add roofline estimation of float8 gemm + overhead (#668)
Vasiliy Kuznetsov committed
027bf39