张壹 zhangyiss
0 Followers · 0 Following
  • Joined on 2024-09-10
Repositories
54
 Projects Packages Public Activity Starred Repositories
zhangyiss synced commits to refs/pull/2290/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:14 +08:00
04a523d77e Merge cb4dc59a9e into 85873cb162
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
fb4e8b896b patch bump (#2343)
Compare 5 commits »
zhangyiss synced commits to refs/pull/2300/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:14 +08:00
d82367b41b Merge cc4de6a607 into 85873cb162
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
Compare 4 commits »
zhangyiss synced commits to fix-type-promotion at zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
zhangyiss synced new reference fix-type-promotion to zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
zhangyiss synced commits to main at zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
zhangyiss synced commits to refs/pull/1970/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
e089584c52 Merge 688e421184 into 85873cb162
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
Compare 2 commits »
zhangyiss synced commits to refs/pull/2219/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
915cc0e531 Merge d2e0b0465c into 85873cb162
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
fb4e8b896b patch bump (#2343)
Compare 19 commits »
zhangyiss synced commits to refs/pull/2234/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:13 +08:00
b25e218ed0 Merge 4d68bd3250 into 85873cb162
85873cb162 [CUDA] Do vectorized store/load in contiguous elementwise ops (#2342)
e14ee12491 add zero for argsort vjp (#2345)
Compare 3 commits »
zhangyiss synced and deleted reference refs/tags/refs/pull/2342/merge at zhangyiss/mlx from mirror 2025-07-10 17:41:12 +08:00
zhangyiss synced commits to refs/pull/1970/merge at zhangyiss/mlx from mirror 2025-07-10 09:31:14 +08:00
d19f3ac486 Merge 688e421184 into e14ee12491
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
fb4e8b896b patch bump (#2343)
2ca533b279 Fix compilation with CUDA 11 (#2331)
Compare 10 commits »
zhangyiss synced commits to refs/pull/2234/merge at zhangyiss/mlx from mirror 2025-07-10 09:31:14 +08:00
7ea2cfc19a Merge 4d68bd3250 into 8b9a3f3cea
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
Compare 2 commits »
zhangyiss synced commits to refs/pull/2342/head at zhangyiss/mlx from mirror 2025-07-10 09:31:14 +08:00
e3534c2db8 Contig uses uint as index and non-contig uses int
zhangyiss synced commits to refs/pull/2342/merge at zhangyiss/mlx from mirror 2025-07-10 09:31:14 +08:00
06fe7db434 Merge e3534c2db8 into e14ee12491
e3534c2db8 Contig uses uint as index and non-contig uses int
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
Compare 4 commits »
zhangyiss synced and deleted reference refs/tags/refs/pull/2339/merge at zhangyiss/mlx from mirror 2025-07-10 09:31:13 +08:00
zhangyiss synced commits to main at zhangyiss/mlx from mirror 2025-07-10 09:31:13 +08:00
e14ee12491 add zero for argsort vjp (#2345)
8b9a3f3cea Align mlx::core::max op nan propagation with NumPy (#2339)
Compare 2 commits »
zhangyiss synced new reference refs/tags/v0.26.3 to zhangyiss/mlx from mirror 2025-07-09 17:14:59 +08:00
zhangyiss synced commits to refs/pull/2339/merge at zhangyiss/mlx from mirror 2025-07-09 17:14:58 +08:00
cde0f3e784 Merge 9a742090ae into fb4e8b896b
9a742090ae Remove tuple unpacking syntax to comply with earlier python versions. Add cuda skip to nanpropagation tests, fix cuda implementation in a separate PR.
aca7fac9ef Make the max nanpropagation test more meaningful for integer types
8b15773206 Add cpu Max nanpropagation. Fix a small fib in cpu max dispatch data types for int8/int16.
fb4e8b896b patch bump (#2343)
Compare 5 commits »
zhangyiss synced commits to refs/pull/2342/head at zhangyiss/mlx from mirror 2025-07-09 17:14:58 +08:00
5c932c7bb0 Use uint as index type
5c3663d4a7 Fix tests on large arrays
e66b685a08 binary => binary_two in binary_two.cu
70ade3015f Use int32_t for IdxT
5459b54bcd Do vectorized store/load in ternary ops
Compare 11 commits »
zhangyiss synced commits to refs/pull/2342/merge at zhangyiss/mlx from mirror 2025-07-09 17:14:58 +08:00
d6d4ffc751 Merge 5c932c7bb0 into fb4e8b896b
5c932c7bb0 Use uint as index type
5c3663d4a7 Fix tests on large arrays
e66b685a08 binary => binary_two in binary_two.cu
70ade3015f Use int32_t for IdxT
Compare 10 commits »
zhangyiss synced commits to refs/tags/v0.26.3 at zhangyiss/mlx from mirror 2025-07-09 17:14:58 +08:00