Code

Local execution e2e training#1472

Code

Open

reyna-abhyankar wants to merge 40 commits intoflexflow:masterflexflow/flexflow-train:masterfrom reyna-abhyankar:local-e2e-trainingreyna-abhyankar/FlexFlow:local-e2e-trainingCopy head branch name to clipboard

+4,462-1,447

Commits

Commits on Aug 25, 2024

• temporary weight adjust index

  reyna-abhyankar

  committed
  6adb290

  Copy full SHA for 6adb290

Commits on Aug 27, 2024

• Loss function

  reyna-abhyankar

  committed
  61697c2

  Copy full SHA for 61697c2

• Add cuda test for loss function

  reyna-abhyankar

  committed
  b56c046

  Copy full SHA for b56c046

• Format

  reyna-abhyankar

  committed
  f75a3d4

  Copy full SHA for f75a3d4

• Refactor and build optimizer kernels, op

  reyna-abhyankar

  committed
  f74711f

  Copy full SHA for f74711f

• Finish optimizer local backing

  reyna-abhyankar

  committed
  40c6252

  Copy full SHA for 40c6252

• Format

  reyna-abhyankar

  committed
  ad9b9ea

  Copy full SHA for ad9b9ea

• E2E update test

  reyna-abhyankar

  committed
  1ddfade

  Copy full SHA for 1ddfade

• Format

  reyna-abhyankar

  committed
  dde9496

  Copy full SHA for dde9496

Commits on Sep 11, 2024

• Small fixes

  reyna-abhyankar

  committed
  59635d8

  Copy full SHA for 59635d8

• Format

  reyna-abhyankar

  committed
  103ef07

  Copy full SHA for 103ef07

Commits on Sep 18, 2024

• Fix test and small issues

  reyna-abhyankar

  committed
  f48f9ff

  Copy full SHA for f48f9ff

• Format

  reyna-abhyankar

  committed
  189c9c8

  Copy full SHA for 189c9c8

Commits on Oct 1, 2024

• Merge branch 'repo-refactor' into local-e2e-training

  reyna-abhyankar

  committed
  d93f464

  Copy full SHA for d93f464

• Pass tests after merge

  reyna-abhyankar

  committed
  b5647c8

  Copy full SHA for b5647c8

• Fix input/weight differentiation

  reyna-abhyankar

  committed
  f5ff91e

  Copy full SHA for f5ff91e

• Fix signature to use unified rep

  reyna-abhyankar

  committed
  7470e71

  Copy full SHA for 7470e71

• Fix model training instance abstraction

  reyna-abhyankar

  committed
  deece1b

  Copy full SHA for deece1b

• Change subcase test name

  reyna-abhyankar

  committed
  1d3cc94

  Copy full SHA for 1d3cc94

Commits on Oct 16, 2024

• Quick fixes

  reyna-abhyankar

  committed
  3cf5d08

  Copy full SHA for 3cf5d08

Commits on Oct 22, 2024

• Refactor training backing and instance

  reyna-abhyankar

  committed
  79ef4c9

  Copy full SHA for 79ef4c9

Commits on Nov 13, 2024

• Expose op folders publicly

  reyna-abhyankar

  committed
  a73b1c3

  Copy full SHA for a73b1c3

• Add tensor type, operate over reduced tensor

  reyna-abhyankar

  committed
  c6fed29

  Copy full SHA for c6fed29

Commits on Jan 7, 2025

• Fixes

  reyna-abhyankar

  committed
  0cdfb1a

  Copy full SHA for 0cdfb1a

Commits on Jan 15, 2025

• Remove tensor lower

  reyna-abhyankar

  committed
  9d252b3

  Copy full SHA for 9d252b3

Commits on Jan 17, 2025

• Add tensor and task lowering scheme

  reyna-abhyankar

  committed
  895c117

  Copy full SHA for 895c117

Commits on Jan 22, 2025

• Build local exec

  reyna-abhyankar

  committed
  411017d

  Copy full SHA for 411017d

Commits on Feb 1, 2025

• Disaggregate local backend

  reyna-abhyankar

  committed
  0128abb

  Copy full SHA for 0128abb

• Update task binding interface and cost estimator

  reyna-abhyankar

  committed
  277f8c2

  Copy full SHA for 277f8c2

Commits on Feb 4, 2025

• Merge master into local execution

  reyna-abhyankar

  committed
  377c6aa

  Copy full SHA for 377c6aa

Commits on Feb 6, 2025

• Build

  reyna-abhyankar

  committed
  8efaec7

  Copy full SHA for 8efaec7

• Format

  reyna-abhyankar

  committed
  1dc1398

  Copy full SHA for 1dc1398

• Split task spec files

  reyna-abhyankar

  committed
  17ad5c8

  Copy full SHA for 17ad5c8

• Delete outdated sim environment file

  reyna-abhyankar

  committed
  639c2c1

  Copy full SHA for 639c2c1

Commits on Feb 13, 2025

• Finish API

  reyna-abhyankar

  committed
  a697044

  Copy full SHA for a697044

• Add tests for allocated and unallocated

  reyna-abhyankar

  committed
  187a8d5

  Copy full SHA for 187a8d5

• Fix nonnegative

  reyna-abhyankar

  committed
  a0f8113

  Copy full SHA for a0f8113

• Format

  reyna-abhyankar

  committed
  b1eab94

  Copy full SHA for b1eab94

• Pass allocated-unallocated tests

  reyna-abhyankar

  committed
  b532c50

  Copy full SHA for b532c50

• Update task registry tests

  reyna-abhyankar

  committed
  f28e5c2

  Copy full SHA for f28e5c2