COMMITS
/ torchao/float8/float8_ops.py July 23, 2025
V
fix autocast handling for float8 training rowwise recipes (#2587)
Vasiliy Kuznetsov committed
July 12, 2025
J
Rename torchao.float8.Float8Tensor to Float8TrainingTensor (#2535)
Jerry Zhang committed
July 7, 2025
Y
test rowwise fp32
y-sq committed
June 2, 2025
D
[float8 training] remove duplicate override for view (#2269)
Daniel Vega-Myhre committed
April 22, 2025
J
Update ruff version to 0.11.6 (#2103)
Jerry Zhang committed
March 14, 2025
D
[float8] Make check for rowwise scales more explicit/readable (#1890)
Daniel Vega-Myhre committed
March 7, 2025
M
fix typing annotation (#1849)
Mayank Mishra committed
February 26, 2025
V
[1/x] float8 cleanup: remove float8_python_api (#1779)
Vasiliy Kuznetsov committed
February 18, 2025
V
fix tensor parallelism for float8 training with rowwise scaling (#1718)
Vasiliy Kuznetsov committed
December 2, 2024
M
check `scale.ndim` before applying `t`/`transpose` (#1339)
Masaki Kozuki committed
November 11, 2024
A
Ruff Lint fixes (torchao/float8) (#1239)
Apurva Jain committed
October 24, 2024
A
Move float8_aten_api to float8_ops (#1155)
Apurva Jain committed
October 7, 2024
V
float8 training axiswise scaling support with per-gemm-argument configuration (#940)
Vasiliy Kuznetsov committed
V
add axiswise scaling to Float8Linear (#920)
Vasiliy Kuznetsov committed
V
add axiswise granularity to Float8Tensor (#919)
Vasiliy Kuznetsov committed
August 12, 2024
V
fix bug with float8 + inference_mode (#659)
Vasiliy Kuznetsov committed
July 30, 2024
V
move float8_experimental to torchao/float8
Vasiliy Kuznetsov committed