Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Apr 1, 2024

• runpod integration

  pythonlearner1025

  committed
  e0cb723

  Copy full SHA for e0cb723

Commits on Mar 28, 2024

• functioning script locally

  pythonlearner1025

  committed
  77b3366

  Copy full SHA for 77b3366

• nvidia working - follow instructions https://bbs.archlinux.org/viewtopic.php?id=266915 and set no-cgroups = false

  pythonlearner1025

  committed
  e41962f

  Copy full SHA for e41962f

• latest

  pythonlearner1025

  committed
  578a906

  Copy full SHA for 578a906

Commits on Mar 24, 2024

• trainer for dep with comfyui

  pythonlearner1025

  committed
  9c17c99

  Copy full SHA for 9c17c99

Commits on Mar 7, 2024

• indicate training complete

  pythonlearner1025

  committed
  2cf1f95

  Copy full SHA for 2cf1f95

Commits on Mar 6, 2024

• append path before import modules

  pythonlearner1025

  committed
  7b2d0d2

  Copy full SHA for 7b2d0d2

Commits on Mar 3, 2024

• fix stable cascade lora key mapping

  Nerogar

  committed
  6ed0b0b

  Copy full SHA for 6ed0b0b

• move dropout setting to the LoRA tab

  Nerogar

  committed
  8195181

  Copy full SHA for 8195181

• Merge branch 'master' into fork/dropout

  Show description for 461ebd6
  Nerogar

  committed
  461ebd6

  Copy full SHA for 461ebd6

• cleanup LoRA dropout

  Nerogar

  committed
  c6fe55d

  Copy full SHA for c6fe55d

• Merge branch 'stochastic_rounding'

  Nerogar

  committed
  5172004

  Copy full SHA for 5172004

• add a disclaimer to the modified optimizer step functions

  Nerogar

  committed
  df100f4

  Copy full SHA for df100f4

• only enable stochastic rounding for bf16 weights

  Nerogar

  committed
  f5852cc

  Copy full SHA for f5852cc

Commits on Mar 2, 2024

• fix loading of samples and concepts when loading a packed config file

  Nerogar

  committed
  b1c2f08

  Copy full SHA for b1c2f08

• add stochastic rounding as an optimizer setting to adam, adamw and adafactor

  Nerogar

  committed
  c2b800a

  Copy full SHA for c2b800a

• Remove dropout setting from the constructor.

  Show description for f92489e
  mx

  committed
  f92489e

  Copy full SHA for f92489e

• improve the efficiency of the stochastic copy function

  Nerogar

  committed
  43fae48

  Copy full SHA for 43fae48

• dropout_pct -> dropout_probability

  mx

  committed
  cd2f232

  Copy full SHA for cd2f232

Commits on Mar 1, 2024

• Merge branch 'master' into stochastic_rounding

  Nerogar

  committed
  c4b7330

  Copy full SHA for c4b7330

• fix a ui exception when switching to a different preset

  Nerogar

  committed
  311b739

  Copy full SHA for 311b739

• fix early stopping of text encoder training

  Nerogar

  committed
  0698f76

  Copy full SHA for 0698f76

• Merge branch 'master' into stochastic_rounding

  Nerogar

  committed
  83ab7e3

  Copy full SHA for 83ab7e3

• fix a model loader bug when training embeddings

  Nerogar

  committed
  30149fc

  Copy full SHA for 30149fc

• fix SD lora training

  Nerogar

  committed
  d25142c

  Copy full SHA for d25142c

Commits on Feb 29, 2024

• experimental stochastic rounding for adafactor and bf16 weights

  Nerogar

  committed
  b47086f

  Copy full SHA for b47086f

• fix ui labels

  Nerogar

  committed
  1c0768d

  Copy full SHA for 1c0768d

• better "stop_training_after" default settings

  Nerogar

  committed
  f34c101

  Copy full SHA for f34c101

Commits on Feb 28, 2024

• fix early stopping of model part training

  Nerogar

  committed
  112a0e1

  Copy full SHA for 112a0e1

• Merge branch 'master' into model_part_config

  Show description for d4e382c
  Nerogar

  committed
  d4e382c

  Copy full SHA for d4e382c

Commits on Feb 24, 2024

• Fix with less typos...

  mx

  committed
  e937e8e

  Copy full SHA for e937e8e

• Fix dropout in state dict (forgot everything had to be a tensor).

  mx

  committed
  02f3843

  Copy full SHA for 02f3843

Commits on Feb 23, 2024

• Add dropout for lora training.

  mx

  committed
  85c6947

  Copy full SHA for 85c6947

• fix stable cascade attention lora keys

  Nerogar

  committed
  93b935e

  Copy full SHA for 93b935e

Commits on Feb 22, 2024

• model spec header for stable cascade lora and embedding files

  Nerogar

  committed
  19e2601

  Copy full SHA for 19e2601