Commits

Breadcrumbs

History for

onremove-sentence-in-learnable-quantization-documentation

User selector

All users

Datepicker

All time

Commit History

Commits on May 26, 2024

• [decomp] Fix default values missing from inplace `rrelu` decomposition (#126978)

  Show description for 39de628
  

  peterbell10

  authored and

  pytorchmergebot

  committed
  39de628

  Copy full SHA for 39de628

Commits on May 7, 2024

• Enable UFMT on `test_decomp.py`, `test_expanded_weights.py` and some files (#125117)

  Show description for c165a8e
  

  shink

  authored and

  pytorchmergebot

  committed
  c165a8e

  Copy full SHA for c165a8e

Commits on May 1, 2024

• [privateuse1] _refs.masked_fill support privateuse1 when value.device.type is cpu (#124835)

  Show description for af67704
  

  bmeoaountiful

  authored and

  pytorchmergebot

  committed
  af67704

  Copy full SHA for af67704

Commits on Apr 20, 2024

• Fix test_decomp test for ops with py_impl(CompositeImplicitAutograd) (#116832)

  Show description for 97ccfad
  

  isuruf

  authored and

  pytorchmergebot

  committed
  97ccfad

  Copy full SHA for 97ccfad

Commits on Apr 11, 2024

• Fixed arange decomp for float dtype (#121013)

  Show description for 6b77415
  

  vfdev-5

  authored and

  pytorchmergebot

  committed
  6b77415

  Copy full SHA for 6b77415

Commits on Apr 5, 2024

• [dynamo, 3.12] enable tests disabled due to missing dynamo 3.12 support (#123300)

  Show description for cbde0f0
  

  williamwen42

  authored and

  pytorchmergebot

  committed
  cbde0f0

  Copy full SHA for cbde0f0

Commits on Feb 14, 2024

• Add linux cpu test for 3.12 (#117853)

  Show description for 244b124
  

  atalman

  authored and

  pytorchmergebot

  committed
  244b124

  Copy full SHA for 244b124

Commits on Feb 6, 2024

• [decomp] Add tests for different dtypes to SDPA decomposition (#119239)

  Show description for 1e4b408
  

  larryliu0820

  authored and

  pytorchmergebot

  committed
  1e4b408

  Copy full SHA for 1e4b408

Commits on Jan 31, 2024

• Fix missing default dim param in weight norm interface decomp (#118762)

  Show description for e87ac82
  

  eellison

  authored and

  pytorchmergebot

  committed
  e87ac82

  Copy full SHA for e87ac82

Commits on Jan 17, 2024

• [PyTorch] SDPA decomp: actually use attn_mask (#117579)

  Show description for e2830e6
  

  digantdesai

  authored and

  pytorchmergebot

  committed
  e2830e6

  Copy full SHA for e2830e6

Commits on Jan 12, 2024

• Add default parameters to rrelu_with_noise() (#117141)

  Show description for 638f85f
  

  aorenste

  authored and

  pytorchmergebot

  committed
  638f85f

  Copy full SHA for 638f85f

Commits on Jan 10, 2024

• [export] Modify SDPA decomposition to decompose _scaled_dot_product_flash_attention_for_cpu (#117097)

  Show description for 8783fe9
  

  larryliu0820

  authored and

  pytorchmergebot

  committed
  8783fe9

  Copy full SHA for 8783fe9

• Fix 0-dim Index in Index Copy decomp (#117065)

  Show description for d654003
  

  eellison

  authored and

  pytorchmergebot

  committed
  d654003

  Copy full SHA for d654003

Commits on Dec 15, 2023

• unMarkDynamoStrictTest on OpInfo-based tests (#115856)

  Show description for 3477a2e
  

  zou3519

  authored and

  pytorchmergebot

  committed
  3477a2e

  Copy full SHA for 3477a2e

Commits on Nov 29, 2023

• Deprecate primTorch module, replace it with decompositions in module Owners (#114754)

  Show description for ba4285b
  

  atalman

  authored and

  pytorchmergebot

  committed
  ba4285b

  Copy full SHA for ba4285b

Commits on Nov 8, 2023

• [decomp] Fix _scaled_dot_product_flash_attention decomposition bug (#113102)

  Show description for 5506b9d
  

  larryliu0820

  authored and

  pytorchmergebot

  committed
  5506b9d

  Copy full SHA for 5506b9d

Commits on Nov 1, 2023

• Add a decomposition for _weight_norm_interface. (#112193)

  Show description for 5a6f801
  

  qihqi

  authored and

  pytorchmergebot

  committed
  5a6f801

  Copy full SHA for 5a6f801

Commits on Oct 31, 2023

• Use `pytree.arg_tree_leaves` everywhere (#112394)

  Show description for 66c32d0
  

  peterbell10

  authored and

  pytorchmergebot

  committed
  66c32d0

  Copy full SHA for 66c32d0

Commits on Oct 30, 2023

• Use `pytree.tree_leaves` everywhere (#112324)

  Show description for bbd5b93
  

  peterbell10

  authored and

  pytorchmergebot

  committed
  bbd5b93

  Copy full SHA for bbd5b93

Commits on Oct 19, 2023

• Mark more decomp tests as slow (#111524)

  Show description for 4f0cf1e
  

  malfet

  authored and

  pytorchmergebot

  committed
  4f0cf1e

  Copy full SHA for 4f0cf1e

Commits on Oct 18, 2023

• Skip `test_quick_core_backward_baddbmm_cuda_float64` (#111493)

  Show description for 16cb3bd
  

  malfet

  authored and

  pytorchmergebot

  committed
  16cb3bd

  Copy full SHA for 16cb3bd

Commits on Oct 11, 2023

• Revert "[core ATen IR] Add decompositions for max, min, var_mean (#110906)"

  Show description for 98c329b
  pytorchmergebot

  committed
  98c329b

  Copy full SHA for 98c329b

• [core ATen IR] Add decompositions for max, min, var_mean (#110906)

  Show description for 9606cda
  

  SS-JIA

  authored and

  pytorchmergebot

  committed
  9606cda

  Copy full SHA for 9606cda

Commits on Oct 6, 2023

• [decomp] Fix different return type in threshold_backward vs. eager (#110689)

  Show description for 7cc0020
  

  cdzhan

  authored and

  pytorchmergebot

  committed
  7cc0020

  Copy full SHA for 7cc0020

Commits on Sep 27, 2023

• [core IR] Add glu as a core decomposition (#110043)

  Show description for 9928c10
  

  SS-JIA

  authored and

  pytorchmergebot

  committed
  9928c10

  Copy full SHA for 9928c10

Commits on Sep 11, 2023

• Introduce Tensor overload to linspace and logspace (#104889)

  Show description for b2cba43
  

  qqaatw

  authored and

  pytorchmergebot

  committed
  b2cba43

  Copy full SHA for b2cba43

• Revert "Introduce Tensor overload to linspace and logspace (#104889)"

  Show description for a7f5abe
  pytorchmergebot

  committed
  a7f5abe

  Copy full SHA for a7f5abe

• Introduce Tensor overload to linspace and logspace (#104889)

  Show description for 57e5239
  

  qqaatw

  authored and

  pytorchmergebot

  committed
  57e5239

  Copy full SHA for 57e5239

Commits on Aug 31, 2023

• Allow registering decomps for HigherOrderOp; add decomp for out_dtype (#108080)

  Show description for 0e4752b
  

  zou3519

  authored and

  pytorchmergebot

  committed
  0e4752b

  Copy full SHA for 0e4752b

Commits on Aug 18, 2023

• [pt2] test if core decomps are differentiable (#107241)

  Show description for 77f080e
  

  nkaretnikov

  authored and

  pytorchmergebot

  committed
  77f080e

  Copy full SHA for 77f080e

Commits on Aug 14, 2023

• [Inductor] Make softshrink composite implicit (#107052)

  Show description for 2c5f96d
  

  lezcano

  authored and

  pytorchmergebot

  committed
  2c5f96d

  Copy full SHA for 2c5f96d

• Make hardshrink's decomp composite implicit (#107039)

  Show description for 3b1254e
  

  lezcano

  authored and

  pytorchmergebot

  committed
  3b1254e

  Copy full SHA for 3b1254e

Commits on Aug 11, 2023

• Implement decomposition for aten.rrelu_with_noise (#106812)

  Show description for e165938
  

  masnesral

  authored and

  pytorchmergebot

  committed
  e165938

  Copy full SHA for e165938

Commits on Aug 4, 2023

• fix: nll_loss batch rule with negative ignore_idx (#106118)

  Show description for a899333
  

  kshitij12345

  authored and

  pytorchmergebot

  committed
  a899333

  Copy full SHA for a899333

Commits on Jul 4, 2023

• [decomp] Add test tracking core ATen operators (#104262)

  Show description for 5c580a9
  

  peterbell10

  authored and

  pytorchmergebot

  committed
  5c580a9

  Copy full SHA for 5c580a9

Pagination

Previous

Next