COMMITS
/ torchao/experimental/quant_api.py May 9, 2025
M
metal lowbit kernels: qmv_fast optimization (#2167)
Manuel Candales committed
April 22, 2025
J
Update ruff version to 0.11.6 (#2103)
Jerry Zhang committed
April 15, 2025
S
Remove zero_point_domain from quant configs (#2058)
Scott Roy committed
April 10, 2025
S
Remove old code from torchao.experimental.quant_api (#2030)
Scott Roy committed
April 8, 2025
S
Move Int8DynamicActivationIntxWeightConfig out of experimental (#1968)
Scott Roy committed
S
March 25, 2025
S
Move config out of experimental (#1954)
Scott Roy committed
March 24, 2025
S
Add dynamic shape support for lowbit kernels (#1942)
Scott Roy committed
March 22, 2025
S
Add quant api + python test for shared embedding (#1937)
Scott Roy committed
March 13, 2025
S
Add bias support to torchao kernels (#1879)
Scott Roy committed
March 7, 2025
V
fix ruff on main branch (#1860)
Vasiliy Kuznetsov committed
S
Migrate to config for Int8DynamicActivationIntxWeightConfig (#1836)
Scott Roy committed
January 30, 2025
S
Ruff lint (#1646)
Scott Roy committed
N
[Feat]: Add support for kleidiai quantization schemes (#1447)
Nikhil Gupta committed
January 14, 2025
S
Clean up linear_int8_dynamic_activation_intx_weight_subclass
Scott Roy committed
January 8, 2025
A
Lint ao (#1521)
Apurva Jain committed
December 13, 2024
M
metal lowbit kernels: executorch ops
Manuel Candales committed
November 26, 2024
S
Enable 8-bit (#1254)
Scott Roy committed
November 19, 2024
M
intx weight only linear quantizer for mps
Manuel Candales committed
October 30, 2024
S
Subclass API (#966)
Scott Roy committed
October 23, 2024
C
Introduce 7-bit quantization for Llama in torchchat.
c4lcut3c committed
October 22, 2024
S
Add embedding ops aten
Scott Roy committed
October 18, 2024
S
Move common ET/Aten op stuff to ops/library.h
Scott Roy committed
October 10, 2024
S
Enable 6-bit kernel
Scott Roy committed
September 30, 2024
S
Change arg order in lowbit linear ops to match aten (#982)
Scott Roy committed
September 26, 2024
S
Change n, k, group_size tensors to have no elements
Scott Roy committed
September 25, 2024
S
Add torchchat quantizer
Scott Roy committed