kohya-ss/sd-scripts

Pull Requests Commits

Merge branch 'dev' into gradual_latent_hires_fix

kohya-ss committed 1 year ago

8f6f734a

Merge pull request #1085 from kohya-ss/dev

kohya-ss committed 1 year ago

Verified cd19df49

update README.md

kohya-ss committed 1 year ago

736365bd

Merge branch 'main' into dev

kohya-ss committed 1 year ago

6ceedb94

Merge pull request #1084 from fireicewolf/devel

kohya-ss committed 1 year ago

Verified 930a3912

Merge pull request #1079 from feffy380/fix/fp8savestate

kohya-ss committed 1 year ago

Verified cf790d87

fireicewolf committed 1 year ago

4e67fb84

fireicewolf committed 1 year ago

50f631c7

fireicewolf committed 1 year ago

85bc371e

Update requirements.txt

feffy380 committed 1 year ago

Verified 322ee52c

Fix ControlNetLLLite training issue #1069

kohya-ss committed 1 year ago

c576f806

Refactor device determination to function; add MPS fallback

akx committed 1 year ago

478156b4

Refactor memory cleaning into a single function

akx committed 1 year ago

afc38707

Log accelerator device

akx committed 1 year ago

2e4bee6f

Merge pull request #1067 from kohya-ss/dev

kohya-ss committed 1 year ago

Verified d5ab97b6

kohya-ss committed 1 year ago

7cb44e45

Merge pull request #1064 from KohakuBlueleaf/fix-grad-sync

kohya-ss committed 1 year ago

Verified 7a20df5a

Merge pull request #1060 from akx/refactor-xpu-init

kohya-ss committed 1 year ago

Verified bea4362e

fix TI training crashes in multigpu #1019

kohya-ss committed 1 year ago

6805cafa

Avoid always sync

KohakuBlueleaf committed 1 year ago

711b40cc

Fix dtype issue in PyTorch 2.0 for generating samples in training sdxl network

kohya-ss committed 1 year ago

696dd7f6

kohya-ss committed 1 year ago

e0a3c692

Add options to reduce memory usage in extract_lora_from_models.py closes #1059

kohya-ss committed 1 year ago

c59249a6

Add network_multiplier for dataset and train LoRA

kohya-ss committed 1 year ago

fef17296

format by black

kohya-ss committed 1 year ago

5a1ebc4c

kohya-ss committed 1 year ago

2a0f45ae

fix to work sample generation in fp8 ref #1057

kohya-ss committed 1 year ago

1f77bb6e

fix to work with torch 2.0

kohya-ss committed 1 year ago

a7ef6422

[Experimental Feature] FP8 weight dtype for base model when running train_network (or sdxl_train_network) (#1057)

KohakuBlueleaf committed 1 year ago

Verified 9cfa68c9

Deduplicate ipex initialization code

akx committed 1 year ago

6f3f701d

Newer Older