malfet

Follow

Nikita Shulga malfet

Follow

295 followers · 23 following

California
https://malfet.github.io/

Achievements

Achievements

Popular repositories Loading

llm_experiments Public

Metal 6 3
pocketfft Public

Clone of https://gitlab.mpcdf.mpg.de/mtr/pocketfft

C 2
gpt-fast Public

Forked from pytorch-labs/gpt-fast

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

Python 2 1
oneDNN Public

Forked from uxlfoundation/oneDNN

oneAPI Deep Neural Network Library (oneDNN)

C++ 2
thrust Public

Forked from NVIDIA/thrust

Thrust is a parallel algorithms library which resembles the C++ Standard Template Library (STL).

C++ 1
Mandelbrot Public

C++ 1

4,149 contributions in the last year

Learn how we count contributions

Less

More

Activity overview

Contributed to pytorch/pytorch, pytorch/torchchat, pytorch/test-infra and 47 other repositories

Contribution activity

March 2025

Created 73 commits in 2 repositories

Created a pull request in pytorch/pytorch that received 10 comments

[BE] Remove `onlyCPU` decorator from test_local_scalar_dense

Followup from #145717, not sure why author thinks those tests should be limited to one architecture. And fixed similar crashes for CUDA and MPS

+2 −1 lines changed • 10 comments

Opened 52 other pull requests in 1 repository

pytorch/pytorch 51 closed 1 open

[MPS][BE] Move polar/complex to stubs
This contribution was made on Mar 21
[MPS] Replace indexed with strided flavor
This contribution was made on Mar 21
[MPS][BE] Get rid of supports_dense flag
This contribution was made on Mar 21
[MPS][BE] Migrate complex_mul to tensor iterator
This contribution was made on Mar 21
[MPS][BE] Migrate torch.complex to binary_functor
This contribution was made on Mar 21
Do not depend on numpy during the import
This contribution was made on Mar 20
[BE] Introduce lapack_work_to_int function
This contribution was made on Mar 20
[BE] Eliminate TODO for 2022
This contribution was made on Mar 19
[MPSInductor] Move threadfence at the right location
This contribution was made on Mar 18
[Build] Guard per-op headers in ACLUtils.cpp
This contribution was made on Mar 18
[Build] Guard per-op headers inclusion
This contribution was made on Mar 18
Fix broken build within xplat/caffe2
This contribution was made on Mar 18
[MPS] Add bicubic2d_aa
This contribution was made on Mar 18
[EZ][Docker] Remove install_db.sh
This contribution was made on Mar 17
[MPS][BE] Move common binary ops macros to indexing.h
This contribution was made on Mar 15
[EZ][BE] Reuse result_of from c10/metal/utils.h
This contribution was made on Mar 15
[EZ][BE] Remove cross-compilation options from mac-build.yml
This contribution was made on Mar 14
[BE] Parametrize TestMPS.test_binops_dtype_precedence
This contribution was made on Mar 14
[MPS] Fix type promotion for torch.floor_divide
This contribution was made on Mar 14
[EZ] Fix typo in UnaryOps.mm
This contribution was made on Mar 14
[MPSInductor] Add support for atan2
This contribution was made on Mar 14
[MPS] Add support for i1e
This contribution was made on Mar 14
[MPSInductor] Add bessel_[jy][01] ops
This contribution was made on Mar 14
[MPS] Add torch.special.bessel_[jy][01] implementations
This contribution was made on Mar 13
[MPSInductor] Fix argmin/argmax long reductions
This contribution was made on Mar 12
Some pull requests not shown.

Reviewed 136 pull requests in 7 repositories

pytorch/pytorch 25 pull requests

Remove outdated instructions from CI scripts
This contribution was made on Mar 22
[MPS/inductor] Add support for modified_scaled_bessel_k{0,1}
This contribution was made on Mar 22
load_inline no_implicit_headers mode
This contribution was made on Mar 22
[MPS] Add support for scaled_modified_bessel_k1 to eager.
This contribution was made on Mar 22
[MPS] Replace indexed with strided flavor
This contribution was made on Mar 21
Fix broken LazyLinear init
This contribution was made on Mar 21
[MPS] Add support for scaled_modified_bessel_k0 for eager.
This contribution was made on Mar 21
[MPS] Add inline to function definition.
This contribution was made on Mar 21
Remove custom kwargs before calling BuildExtension.__init__(...)
This contribution was made on Mar 21
Update Doc for Intel XPU Profiling
This contribution was made on Mar 21
[MPS] Add support for modified_bessel_k1 to eager and inductor.
This contribution was made on Mar 21
Do not depend on numpy during the import
This contribution was made on Mar 21
Fix clang-tidy errors
This contribution was made on Mar 20
changing linear layer initialization formula in docs
This contribution was made on Mar 20
Specify the default PyTorch Distributed backend for MPS
This contribution was made on Mar 20
Add overflow check for large storage_offsets
This contribution was made on Mar 20
Add triton as dependency to CUDA aarch64 build
This contribution was made on Mar 20
add grad_output shape check for adaptive_avg_pool2d_backward
This contribution was made on Mar 20
[MPS] nanmedian implementation
This contribution was made on Mar 20
Update the heuristic for AArch64 bmm/baddbmm
This contribution was made on Mar 19
Update nightly s390x builds
This contribution was made on Mar 19
xpu: improve error handling and reporting in XPU cmake files
This contribution was made on Mar 19
Add x86-simd-sort accelerated sorting
This contribution was made on Mar 19
Remove Ubuntu 18.04 scripts
This contribution was made on Mar 19
Pin auditwheel to 6.2.0
This contribution was made on Mar 19
Some pull request reviews not shown.

pytorch/test-infra 18 pull requests

Enable torch.compile validation on Amazon 2023
This contribution was made on Mar 21
[viable/strict] Fix failing jobs with unstable issues
This contribution was made on Mar 19
fix dropdown gap causing issues dropping down
This contribution was made on Mar 18
[validations] Remove libtorch PRE_CXX11_ABI from matrix.
This contribution was made on Mar 17
fix width of time column on main hud page
This contribution was made on Mar 17
fix flickering white on page load / reload
This contribution was made on Mar 17
fix monsterize layout - width of columns
This contribution was made on Mar 17
better dark mode colors for highlighting
This contribution was made on Mar 17
fix tooltip width
This contribution was made on Mar 17
styling & CSS Improvements: Add CSS variables for better color consistency
This contribution was made on Mar 14
Temporarily exclude test channel from sha256
This contribution was made on Mar 13
Remove cuda 12.6 aarch
This contribution was made on Mar 13
Cherry-Pick of #6401
This contribution was made on Mar 13
CSS Layout Improvements (extracted from PR #6392)
This contribution was made on Mar 13
HUD dark mode
This contribution was made on Mar 12
monsterized failures in grouped view
This contribution was made on Mar 12
[viable/strict] Ignore failing jobs with unstable issues
This contribution was made on Mar 5
Upgrade github ubuntu-20.04 runners to ubuntu-24.04
This contribution was made on Mar 4

janeyx99/torch-release-notes 1 pull request

Remove example miscategorized PR I've already categorized
This contribution was made on Mar 20

pytorch/benchmark 1 pull request

[Fix Syn] add missing yaml file manually
This contribution was made on Mar 19

pytorch/tutorials 1 pull request

Add OOV token handling to character-level RNN tutorial
This contribution was made on Mar 14

pytorch/executorch 1 pull request

Refactor gather_test_model to be hard coded directly
This contribution was made on Mar 13

triton-lang/triton 1 pull request

Add LLVM aarch64 Almalinux build
This contribution was made on Mar 4

Created an issue in pytorch/pytorch that received 11 comments

UNSTABLE trunk / libtorch-linux-focal-cuda12.4-py3.10-gcc9-debug / build

See https://hud.pytorch.org/hud/pytorch/pytorch/c677f3251f46b4bffdaa7758fb7102d665b6f11b/1?per_page=50&name_filter=%20libtorch-linux-focal-cuda12.4…

11 comments

Opened 7 other issues in 1 repository

pytorch/pytorch 6 open 1 closed

UNSTABLE pull / cuda12.4-py3.10-gcc9-sm75 / test (pr_time_benchmarks)
This contribution was made on Mar 18
☂️ MPS support for large tensors
This contribution was made on Mar 17
[RelEng] wheel testing for new arch versions
This contribution was made on Mar 14
[BE] Combine windows_arm64_binary_build.yaml.js2 with regular windows yaml
This contribution was made on Mar 7
on-pr docker build stuck with user is not authorized to BatchGetImage
This contribution was made on Mar 7
magma builds should be part of the docker image builds
This contribution was made on Mar 7
UNSTABLE trunk / libtorch-linux-focal-cuda12.6-py3.10-gcc9-debug / build
This contribution was made on Mar 7

9 contributions in private repositories Mar 12 – Mar 21