Commits

Branch selector

 main

User selector

andrewor14

Datepicker

All time

Commit History

Commits on May 17, 2024

• [quant][pt2e] Allow multi users without output observers (pytorch#126487)

  Show description for 6931f78
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  6931f78

  Copy full SHA for 6931f78

Commits on May 14, 2024

• [quant] Make per_group and per_token quant match torch.fake_quantize (pytorch#125781)

  Show description for 3cba50e
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  3cba50e

  Copy full SHA for 3cba50e

Commits on Apr 30, 2024

• [quant][pt2e] Fix conv-bn weight + bias per channel QAT (pytorch#125208)

  Show description for 8242fb6
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  8242fb6

  Copy full SHA for 8242fb6

Commits on Apr 24, 2024

• [quant][pt2e] Move batch norm op between eval/train for cuda (pytorch#123957)

  Show description for 85b28ff
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  85b28ff

  Copy full SHA for 85b28ff

• [quant][pt2e] Move batch norm op between eval/train for cuda (pytorch#123957)

  Show description for 4efb28c
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  4efb28c

  Copy full SHA for 4efb28c

Commits on Apr 16, 2024

• [quant] Do not decompose choose_qparams_per_token_asymmetric (pytorch#124178)

  Show description for 3eea300
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  3eea300

  Copy full SHA for 3eea300

Commits on Apr 12, 2024

• [quant] Enable backward for choose_qparams_per_token_asymmetric (pytorch#123452)

  Show description for 762e196
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  762e196

  Copy full SHA for 762e196

• [pt2e][qat] Support conv-transpose-bn[-relu] QAT fusion (pytorch#123652)

  Show description for 5c0a380
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  5c0a380

  Copy full SHA for 5c0a380

Commits on Apr 11, 2024

• [quant] Enable backward for choose_qparams_per_token_asymmetric (pytorch#123452)

  Show description for c839008
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  c839008

  Copy full SHA for c839008

Commits on Apr 3, 2024

• [quant][be] Simplify fake_quant_per_channel (pytorch#123186)

  Show description for fe29a8f
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  fe29a8f

  Copy full SHA for fe29a8f

Commits on Mar 22, 2024

• [quant][pt2] Fix create FQ with FixedQParamsQSpec (pytorch#122104)

  Show description for ea8e0c7
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  ea8e0c7

  Copy full SHA for ea8e0c7

• [jit] Fix _batch_norm_with_update shape function (pytorch#122430)

  Show description for 6e6891e
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  6e6891e

  Copy full SHA for 6e6891e

Commits on Mar 18, 2024

• Batch Norm Consolidation (pytorch#116092)

  Show description for 773ae81
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  773ae81

  Copy full SHA for 773ae81

Commits on Mar 8, 2024

• Batch Norm Consolidation (pytorch#116092)

  Show description for 7b4f70e
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  7b4f70e

  Copy full SHA for 7b4f70e

Commits on Feb 28, 2024

• [quant][pt2e] Relax model_is_exported input (pytorch#120720)

  Show description for 91190d8
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  91190d8

  Copy full SHA for 91190d8

Commits on Feb 16, 2024

• [quant][pt2e] Add model_is_exported util function (pytorch#119726)

  Show description for 6ea4480
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  6ea4480

  Copy full SHA for 6ea4480

Commits on Feb 13, 2024

• [quant][pt2e][be] Rename eval_utils -> export_utils (pytorch#119725)

  Show description for 8ec8d78
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  8ec8d78

  Copy full SHA for 8ec8d78

• [quant][pt2] Fix _disallow_eval_train error message (pytorch#119694)

  Show description for 830ed6d
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  830ed6d

  Copy full SHA for 830ed6d

Commits on Feb 6, 2024

• [quant][pt2e] Allow users to override train/eval behavior (pytorch#119091)

  Show description for 6c1cca1
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  6c1cca1

  Copy full SHA for 6c1cca1

Commits on Feb 2, 2024

• [quant][pt2] Add move_exported_model_to_train (pytorch#113492)

  Show description for 70605d1
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  70605d1

  Copy full SHA for 70605d1

Commits on Dec 12, 2023

• [quant][fx] Lower operator.matmul in convert_fx (pytorch#113954)

  Show description for 6988e40
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  6988e40

  Copy full SHA for 6988e40

Commits on Nov 17, 2023

• [quant][pt2] Support conv1d-bn QAT fusion (pytorch#113714)

  Show description for e5102cc
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  e5102cc

  Copy full SHA for e5102cc

Commits on Nov 16, 2023

• [quant][pt2][be] Rewrite QAT annotations using subgraph matcher (pytorch#113709)

  Show description for 8241fe6
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  8241fe6

  Copy full SHA for 8241fe6

Commits on Nov 15, 2023

• [quant][pt2][be] Refactor QAT tests for future patterns (pytorch#113658)

  Show description for f9ea697
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  f9ea697

  Copy full SHA for f9ea697

Commits on Nov 14, 2023

• [quant][pt2][be] Remove add/relu from conv-bn QAT pattern (pytorch#113006)

  Show description for 14eb92c
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  14eb92c

  Copy full SHA for 14eb92c

Commits on Nov 7, 2023

• [quant][pt2] Fix custom dtype per channel weight in QAT (pytorch#112612)

  Show description for c0aba9b
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  c0aba9b

  Copy full SHA for c0aba9b

Commits on Nov 6, 2023

• [quant][pt2] Support quantized conv bias in QAT fusion (pytorch#112528)

  Show description for b6e85eb
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  b6e85eb

  Copy full SHA for b6e85eb

Commits on Oct 31, 2023

• [quant][pt2][be] Refactor QAT q-dq patterns (pytorch#112279)

  Show description for 5cd1208
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  5cd1208

  Copy full SHA for 5cd1208

• [quant][pt2] Fix QAT conv-bn bias derived qspec (pytorch#112159)

  Show description for 231129e
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  231129e

  Copy full SHA for 231129e

Commits on Oct 20, 2023

• [pt2][quant] Clean up QAT get conv-bn-relu nodes (pytorch#111515)

  Show description for e4e7d34
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  e4e7d34

  Copy full SHA for e4e7d34

Commits on Oct 11, 2023

• [quant][pt2] Preserve source_fn_stack after QAT fusion (pytorch#110899)

  Show description for 0e551bb
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  0e551bb

  Copy full SHA for 0e551bb

Commits on Oct 5, 2023

• Back out "Enable pickling model prepared with QAT qconfig" (pytorch#110392)

  Show description for 7c72238
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  7c72238

  Copy full SHA for 7c72238

• [quant][pt2] Support cudnn_batch_norm in QAT fusion (pytorch#109908)

  Show description for 62cad5b
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  62cad5b

  Copy full SHA for 62cad5b

Commits on Sep 15, 2023

• [quant][be] Move QAT tests to its own file (pytorch#108061)

  Show description for 7da3c93
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  7da3c93

  Copy full SHA for 7da3c93

Commits on Sep 11, 2023

• [quant][pt2] Fix and rename move_model_to_eval (pytorch#108891)

  Show description for e8a402c
  

  andrewor14

  authored and

  pytorchmergebot

  committed
  e8a402c

  Copy full SHA for e8a402c

Pagination

Previous

Next