Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 27, 2024

• [Pytorch][AO] Update choose_qparams_per_token op to output correct shape for scales and zp (pytorch#136807)

  Show description for e5a5793
  

  kimishpatel

  authored and

  pytorchmergebot

  committed
  e5a5793

  Copy full SHA for e5a5793

• [export] simplify automatic dynamic shapes processing (pytorch#136591)

  Show description for 6075f56
  

  pianpwk

  authored and

  pytorchmergebot

  committed
  6075f56

  Copy full SHA for 6075f56

• add types to _dynamo/code_context.py (pytorch#136665)

  Show description for a8b5adc
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  a8b5adc

  Copy full SHA for a8b5adc

• Revert "[user triton] Make tl.constexpr specialization work for triton_op & capture_triton (pytorch#136686)"

  Show description for 287dc36
  pytorchmergebot

  committed
  287dc36

  Copy full SHA for 287dc36

• Fix RMSNorm doc per pytorch#136597 (pytorch#136727)

  Show description for 2208ff6
  

  mikaylagawarecki

  authored and

  pytorchmergebot

  committed
  2208ff6

  Copy full SHA for 2208ff6

• [dynamo] attempt run only mode when dynamo cache limit is hit (pytorch#136655)

  Show description for 2157e39
  

  williamwen42

  authored and

  pytorchmergebot

  committed
  2157e39

  Copy full SHA for 2157e39

• Revert "Add Triton CPU as an Inductor backend (pytorch#133408)"

  Show description for 36428f9
  pytorchmergebot

  committed
  36428f9

  Copy full SHA for 36428f9

• Delete project.default_flavors_mode buckconfig (pytorch#136772)

  Show description for 17f396b
  

  8Keep

  authored and

  pytorchmergebot

  committed
  17f396b

  Copy full SHA for 17f396b

• Remove problematic constructor (pytorch#136708)

  Show description for cbc182d
  

  cyyever

  authored and

  pytorchmergebot

  committed
  cbc182d

  Copy full SHA for cbc182d

• [AOTAutogradCache] Log time taken_ns (pytorch#136529)

  Show description for dc8c0aa
  

  jamesjwu

  authored and

  pytorchmergebot

  committed
  dc8c0aa

  Copy full SHA for dc8c0aa

• [user triton] Make tl.constexpr specialization work for triton_op & capture_triton (pytorch#136686)

  Show description for 9f5b97a
  

  davidberard98

  authored and

  pytorchmergebot

  committed
  9f5b97a

  Copy full SHA for 9f5b97a

• Add a nightly hotpatch utils for python only PR (pytorch#136535)

  Show description for ad51995
  

  bhack

  authored and

  pytorchmergebot

  committed
  ad51995

  Copy full SHA for ad51995

• [MPS] Fix AvgPool2d for float16 (pytorch#136822)

  Show description for 9d72f74
  

  malfet

  authored and

  pytorchmergebot

  committed
  9d72f74

  Copy full SHA for 9d72f74

• [BE][MPS] Delete MacOS12 low-precision ops (pytorch#136821)

  Show description for 2b6f4e9
  

  malfet

  authored and

  pytorchmergebot

  committed
  2b6f4e9

  Copy full SHA for 2b6f4e9

• [inductor] Triton codegen: Use scalar when creating f64 constant instead of 1-element tensor (pytorch#136858)

  Show description for 45a8b56
  

  masnesral

  authored and

  pytorchmergebot

  committed
  45a8b56

  Copy full SHA for 45a8b56

• [aotd] Do not force contiguous() for channels_last (pytorch#135225)

  Show description for 34d788f
  

  IvanKobzarev

  authored and

  pytorchmergebot

  committed
  34d788f

  Copy full SHA for 34d788f

• Revert "Deal with size oblivious before going into worker (pytorch#135137)"

  Show description for de159f0
  pytorchmergebot

  committed
  de159f0

  Copy full SHA for de159f0

• [ONNX] Remove unused functions (pytorch#136609)

  Show description for 1be3d62
  

  justinchuby

  authored and

  pytorchmergebot

  committed
  1be3d62

  Copy full SHA for 1be3d62

• Revert "Don't uselessly recompute axiom dict every static eval call (pytorch#135429)"

  Show description for e5228a7
  pytorchmergebot

  committed
  e5228a7

  Copy full SHA for e5228a7

• Limit number of cores to 16 when benchmarking Inductor on ARM (pytorch#136424)

  Show description for a55aa71
  

  cfRod

  authored and

  pytorchmergebot

  committed
  a55aa71

  Copy full SHA for a55aa71

• Revert "Add deterministic path for CUDA cumsum (pytorch#136224)"

  Show description for e9d2765
  pytorchmergebot

  committed
  e9d2765

  Copy full SHA for e9d2765

• [inductor] [cpp] fix gemm_output_name conflict (pytorch#136419)

  Show description for c2637a7
  

  chunyuan-w

  authored and

  pytorchmergebot

  committed
  c2637a7

  Copy full SHA for c2637a7

• [Flex Attention] fix block size order (pytorch#136657)

  Show description for b42f1e3
  

  BoyuanFeng

  authored and

  pytorchmergebot

  committed
  b42f1e3

  Copy full SHA for b42f1e3

• [aotd] Cleanup on subclasses in inductor freezing (pytorch#136549)

  Show description for 9581508
  

  IvanKobzarev

  authored and

  pytorchmergebot

  committed
  9581508

  Copy full SHA for 9581508

• [Distributed] [13/N] Fix clang-tidy warnings in torch/csrc/distributed/ (pytorch#136713)

  Show description for bbff667
  

  cyyever

  authored and

  pytorchmergebot

  committed
  bbff667

  Copy full SHA for bbff667

• [dynamo] Added support for tensor's is_inference method (pytorch#136450)

  Show description for 48c18ff
  

  SalmanMohammadi

  authored and

  pytorchmergebot

  committed
  48c18ff

  Copy full SHA for 48c18ff

• Using device-agnostic autocast api (pytorch#136613)

  Show description for e14b58f
  

  FFFrog

  authored and

  pytorchmergebot

  committed
  e14b58f

  Copy full SHA for e14b58f

• [PP] Remove modifications to autograd nodes in ZB (pytorch#136678)

  Show description for ad6c70b
  

  H-Huang

  authored and

  pytorchmergebot

  committed
  ad6c70b

  Copy full SHA for ad6c70b

• Refactor offset logic and work for nD (pytorch#135861)

  Show description for 9529d01
  

  zeshengzong

  authored and

  pytorchmergebot

  committed
  9529d01

  Copy full SHA for 9529d01

• [EZ][BE] Add torch.complex to MPS_DTYPES (pytorch#136755)

  Show description for 69bd13d
  

  malfet

  authored and

  pytorchmergebot

  committed
  69bd13d

  Copy full SHA for 69bd13d

• Log total miss inplaced bytes (pytorch#136684)

  Show description for 73f038c
  

  laithsakka

  authored and

  pytorchmergebot

  committed
  73f038c

  Copy full SHA for 73f038c

• Delete grid reduction optimization that is causing specialization (pytorch#136783)

  Show description for 0200bea
  

  oulgen

  authored and

  pytorchmergebot

  committed
  0200bea

  Copy full SHA for 0200bea

• add typing to _dynamo/current_scope_id.py (pytorch#136676)

  Show description for a63d7cb
  

  bobrenjc93

  authored and

  pytorchmergebot

  committed
  a63d7cb

  Copy full SHA for a63d7cb

• Revert "[inductor] Triton codegen: Use scalar when creating f64 constant instead of 1-element tensor (pytorch#136594)"

  Show description for 5eb68d5
  pytorchmergebot

  committed
  5eb68d5

  Copy full SHA for 5eb68d5

• Don't uselessly recompute axiom dict every static eval call (pytorch#135429)

  Show description for 507c69e
  

  ezyang

  authored and

  pytorchmergebot

  committed
  507c69e

  Copy full SHA for 507c69e

Pagination

Previous

Next