Commits

Branch selector

 master

User selector

supriyar

Datepicker

All time

Commit History

Commits on Nov 5, 2020

• [quant][pyper] Add support for pruned weights in embedding_bag_byte lookup (pytorch#47329)

  Show description for d8c3b2b
  

  supriyar

  authored and

  facebook-github-bot

  committed
  d8c3b2b

• [quant] Add testing coverage for 4-bit embedding_bag sparse lookup op (pytorch#47328)

  Show description for 433b55b
  

  supriyar

  authored and

  facebook-github-bot

  committed
  433b55b

Commits on Oct 28, 2020

• [quant] Fix flaky test test_histogram_observer_against_reference (pytorch#46957)

  Show description for fc2bd99
  

  supriyar

  authored and

  facebook-github-bot

  committed
  fc2bd99

Commits on Oct 23, 2020

• [quant][fx] EmbeddingBag quantization support (pytorch#46678)

  Show description for 842494a
  

  supriyar

  authored and

  facebook-github-bot

  committed
  842494a

• [quant][fx] Embedding quantization support (pytorch#46677)

  Show description for e34c825
  

  supriyar

  authored and

  facebook-github-bot

  committed
  e34c825

Commits on Oct 13, 2020

• [quant][graph][fix] Set type for GetAttr nodes in remapTypes (pytorch#46250)

  Show description for 95ccf34
  

  supriyar

  authored and

  facebook-github-bot

  committed
  95ccf34

Commits on Oct 8, 2020

• [quant][pyper] Rename the sparse argument for embedding_bag ops (pytorch#46003)

  Show description for 31888b2
  

  supriyar

  authored and

  facebook-github-bot

  committed
  31888b2

• [quant] Set sparse to False for embedding_bag ops in graph mode (pytorch#45997)

  Show description for 8c80ee8
  

  supriyar

  authored and

  facebook-github-bot

  committed
  8c80ee8

Commits on Oct 7, 2020

• [quant] Refactor qembeddingbag to remove duplicate code (pytorch#45881)

  Show description for 1b31ed3
  

  supriyar

  authored and

  facebook-github-bot

  committed
  1b31ed3

• [quant] Support for 4-bit quantized EmbeddingBag module (pytorch#45865)

  Show description for 43dc7ef
  

  supriyar

  authored and

  facebook-github-bot

  committed
  43dc7ef

• [quant] Support 4-bit embedding_bag operators using the dtype quint4x2 (pytorch#45752)

  Show description for 11c3261
  

  supriyar

  authored and

  facebook-github-bot

  committed
  11c3261

• [quant] Add 4-bit embedding_bag prepack/unpack support using quint4x2 (pytorch#45751)

  Show description for 5c283fa
  

  supriyar

  authored and

  facebook-github-bot

  committed
  5c283fa

Commits on Oct 2, 2020

• [quant] PerChannelFloatQParams support for quint4x2 dtype (pytorch#45594)

  Show description for 1a2d3b6
  

  supriyar

  authored and

  facebook-github-bot

  committed
  1a2d3b6

• [quant] creating quint4x2 dtype for quantized tensors (pytorch#44678)

  Show description for 04526a4
  

  supriyar

  authored and

  facebook-github-bot

  committed
  04526a4

Commits on Sep 30, 2020

• [quant] Make choose_qparams_optimized return Tensors to preserve dtype (pytorch#45530)

  Show description for c112e89
  

  supriyar

  authored and

  facebook-github-bot

  committed
  c112e89

Commits on Sep 29, 2020

• [quant] Add quant APIs to save/load observer state_dict (pytorch#44846)

  Show description for 489af4d
  

  supriyar

  authored and

  facebook-github-bot

  committed
  489af4d

Commits on Sep 25, 2020

• [quant][graph] Remove redundant aten::wait calls in the graph (pytorch#45257)

  Show description for a117d96
  

  supriyar

  authored and

  facebook-github-bot

  committed
  a117d96

Commits on Sep 24, 2020

• [quant] Add optimized approach to calculate qparams for qembedding_bag (pytorch#45149)

  Show description for 60665ac
  

  supriyar

  authored and

  facebook-github-bot

  committed
  60665ac

Commits on Sep 23, 2020

• [quant][fx][bug] Fix error in convert step for QAT (pytorch#45050)

  Show description for 7fba30c
  

  supriyar

  authored and

  facebook-github-bot

  committed
  7fba30c

Commits on Sep 17, 2020

• [quant][qat] Ensure fake_quant and observer can be disabled on scriptmodule (pytorch#44773)

  Show description for 1fde54d
  

  supriyar

  authored and

  facebook-github-bot

  committed
  1fde54d

• [quant][fx] Add node name as prefix to observer module name (pytorch#44765)

  Show description for 361b38d
  

  supriyar

  authored and

  facebook-github-bot

  committed
  361b38d

Commits on Sep 16, 2020

• [quant][qat] Ensure observers and fq modules are scriptable (pytorch#44749)

  Show description for 3f512b0
  

  supriyar

  authored and

  facebook-github-bot

  committed
  3f512b0

Commits on Sep 9, 2020

• [quant] Move EmbeddingBag eager quantization to static (pytorch#44217)

  Show description for 646ffd4
  

  supriyar

  authored and

  facebook-github-bot

  committed
  646ffd4

• [quant] Add quantized Embedding module (pytorch#44208)

  Show description for 57b87aa
  

  supriyar

  authored and

  facebook-github-bot

  committed
  57b87aa

• [quant] Support quantization of embedding lookup operators (pytorch#44207)

  Show description for 6013a29
  

  supriyar

  authored and

  facebook-github-bot

  committed
  6013a29

Commits on Sep 5, 2020

• [quant][pyper] Support quantization of ops in fork-wait subgraph (pytorch#44048)

  Show description for 199c73b
  

  supriyar

  authored and

  facebook-github-bot

  committed
  199c73b

• [quant][pyper] make embedding_bag quantization static (pytorch#44008)

  Show description for 164b96c
  

  supriyar

  authored and

  facebook-github-bot

  committed
  164b96c

• [quant] Support aten::embedding_bag quantization in graph mode (pytorch#43989)

  Show description for a0ae416
  

  supriyar

  authored and

  facebook-github-bot

  committed
  a0ae416

Commits on Aug 24, 2020

• [quant] Support set API for EmbeddingBag quantization (pytorch#43433)

  Show description for 284ff04
  

  supriyar

  authored and

  facebook-github-bot

  committed
  284ff04

• [quant] Add benchmarks for quantized embeddingbag module (pytorch#43296)

  Show description for 7024ce8
  

  supriyar

  authored and

  facebook-github-bot

  committed
  7024ce8

Commits on Aug 21, 2020

• [quant] Enable from_float for quantized Embedding_Bag (pytorch#43176)

  Show description for 3293fdf
  

  supriyar

  authored and

  facebook-github-bot

  committed
  3293fdf

• [quant] Make offsets an optional argument (pytorch#43090)

  Show description for b354b42
  

  supriyar

  authored and

  facebook-github-bot

  committed
  b354b42

• [quant] Create nn.quantized.dynamic.EmbeddingBag (pytorch#43088)

  Show description for 4db8ca1
  

  supriyar

  authored and

  facebook-github-bot

  committed
  4db8ca1

Commits on Aug 20, 2020

• [quant] Add benchmakrs for embedding_bag coversion ops (pytorch#43291)

  Show description for 4fc9e95
  

  supriyar

  authored and

  facebook-github-bot

  committed
  4fc9e95

Commits on Aug 14, 2020

• [quant] Add torchbind support for embedding_bag packed weights (pytorch#42881)

  Show description for 66b3382
  

  supriyar

  authored and

  facebook-github-bot

  committed
  66b3382

Pagination

PreviousNext