v22.6.2 #1995

bmaltais · 2024-02-24T15:43:23Z

2024/02/24 (v22.6.2)

Fix issue with Lora Extract GUI
- Fix syntax issue where parameter lora_network_weights is actually called network_weights
Merge sd-scripts v0.8.4 code update
- Fixed a bug that the VRAM usage without Text Encoder training is larger than before in training scripts for LoRA etc (train_network.py, sdxl_train_network.py).
  - Text Encoders were not moved to CPU.
- Fixed typos. Thanks to akx! PR #1053
- The log output has been improved. PR #905 Thanks to shirayu!
  - The log is formatted by default. The rich library is required. Please see Upgrade and update the library.
  - If rich is not installed, the log output will be the same as before.
  - The following options are available in each training script:
  - --console_log_simple option can be used to switch to the previous log output.
  - --console_log_level option can be used to specify the log level. The default is INFO.
  - --console_log_file option can be used to output the log to a file. The default is None (output to the console).
- The sample image generation during multi-GPU training is now done with multiple GPUs. PR #1061 Thanks to DKnight54!
- The support for mps devices is improved. PR #1054 Thanks to akx! If mps device exists instead of CUDA, the mps device is used automatically.
- The --new_conv_rank option to specify the new rank of Conv2d is added to networks/resize_lora.py. PR #1102 Thanks to mgz-dev!
- An option --highvram to disable the optimization for environments with little VRAM is added to the training scripts. If you specify it when there is enough VRAM, the operation will be faster.
  - Currently, only the cache part of latents is optimized.
- The IPEX support is improved. PR #1086 Thanks to Disty0!
- Fixed a bug that svd_merge_lora.py crashes in some cases. PR #1087 Thanks to mgz-dev!
- DyLoRA is fixed to work with SDXL. PR #1126 Thanks to tamlog06!
- The common image generation script gen_img.py for SD 1/2 and SDXL is added. The basic functions are the same as the scripts for SD 1/2 and SDXL, but some new features are added.
  - External scripts to generate prompts can be supported. It can be called with --from_module option. (The documentation will be added later)
  - The normalization method after prompt weighting can be specified with --emb_normalize_mode option. original is the original method, abs is the normalization with the average of the absolute values, none is no normalization.
- Gradual Latent Hires fix is added to each generation script. See here for details.

remove missing import, and remove unused imports

Update IPEX Libs

fix broken import in svd_merge_lora script

Device support improvements (MPS)

Fix IPEX support and add XPU device to device_utils

…1061) * Update train_util.py Modifying to attempt enable multi GPU inference * Update train_util.py additional VRAM checking, refactor check_vram_usage to return string for use with accelerator.print * Update train_network.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py remove sample image debug outputs * Update train_util.py * Update train_util.py * Update train_network.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_network.py * Update train_util.py * Update train_network.py * Update train_network.py * Update train_network.py * Cleanup of debugging outputs * adopt more elegant coding Co-authored-by: Aarni Koskela <[email protected]> * Update train_util.py Fix leftover debugging code attempt to refactor inference into separate function * refactor in function generate_per_device_prompt_list() generation of distributed prompt list * Clean up missing variables * fix syntax error * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * true random sample image generation update code to reinitialize random seed to true random if seed was set * true random sample image generation * simplify per process prompt * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_util.py * Update train_network.py * Update train_network.py * Update train_network.py --------- Co-authored-by: Aarni Koskela <[email protected]>

Signed-off-by: 陳鈞 <[email protected]>

Too big to run on GitHub free runner. Error: no space left on device This reverts commit fc029d7. Signed-off-by: 陳鈞 <[email protected]>

…:3.10-slim - Change the base image from `python:3.10` to `python:3.10-slim` - Expose ports `7860` and `6006` in the Dockerfile Signed-off-by: 陳鈞 <[email protected]>

- Switch to a slim variant of the Python 3.10 Docker image for the build stage. - Introduce a conditional installation of `pillow-simd` replacing `pillow`, specific to x86 architecture. - Add required dependencies for `pillow-simd` installation, along with cleanup commands to remove unnecessary package lists after installation. - Update runtime dependencies by adding `libtcl8.6` and `libtk8.6` to the final Docker image. Signed-off-by: 陳鈞 <[email protected]>

Signed-off-by: 陳鈞 <[email protected]>

Typo fix

Nicer formatting for options.

Fixed unclosed formatting

Update options.md

Optimize the Dockerfile to reduce the size of the image. (50.76 GB to 19.04 GB)

Label clarifications

Fix dylora create_modules error when training sdxl

Resize lora add new rank for conv

v0.8.4

kohya-ss and others added 30 commits November 23, 2023 22:17

add gradual latent

6849546

Update README.md

610566f

Merge branch 'dev' into gradual_latent_hires_fix

2897a89

fix gradual latent cannot be disabled

298c6c2

apply unsharp mask

2c50ea0

add unsharp mask

29b6fa6

fix strength error

2952bca

add target_x flag (not sure this impl is correct)

7a4e507

Merge branch 'dev' into gradual_latent_hires_fix

e8c3a02

Merge branch 'dev' into gradual_latent_hires_fix

9278031

fix controlnet to work with gradual latent

07ef03d

enable comment in prompt file, record raw prompt to metadata

d61ecb2

Merge branch 'dev' into gradual_latent_hires_fix

da9b34f

Log accelerator device

2e4bee6

Refactor memory cleaning into a single function

afc3870

Refactor device determination to function; add MPS fallback

478156b

Merge branch 'dev' into gradual_latent_hires_fix

8f6f734

Update IPEX Libs

ccc3a48

fix broken import in svd_merge_lora script

d4b9568

remove missing import, and remove unused imports

IPEX torch.tensor FP64 workaround

988dee0

Merge pull request #1086 from Disty0/dev

9d7729c

Update IPEX Libs

Merge pull request #1087 from mgz-dev/fix-imports-on-svd_merge_lora

7f948db

fix broken import in svd_merge_lora script

Merge pull request #1054 from akx/mps

2ca4d0c

Device support improvements (MPS)

Fix IPEX support and add XPU device to device_utils

a6a2b5a

Merge pull request #1092 from Disty0/dev_device_support

9f0f0d5

Fix IPEX support and add XPU device to device_utils

add highvram option and do not clear cache in caching latents

5cca1fd

simplify multi-GPU sample generation

11aced3

fix typo

2f9a344

unificaition of gen scripts for SD and SDXL, work in progress

6269682

jim60105 and others added 29 commits February 18, 2024 09:59

ci(docker): Add docker CI

dc94512

Signed-off-by: 陳鈞 <[email protected]>

Revert "ci(docker): Add docker CI"

5dc9db6

Too big to run on GitHub free runner. Error: no space left on device This reverts commit fc029d7. Signed-off-by: 陳鈞 <[email protected]>

chore(docker): Add EXPOSE ports and change final base image to python…

8330597

…:3.10-slim - Change the base image from `python:3.10` to `python:3.10-slim` - Expose ports `7860` and `6006` in the Dockerfile Signed-off-by: 陳鈞 <[email protected]>

chore(docker): Add label

d7add28

Signed-off-by: 陳鈞 <[email protected]>

fix dylora create_modules error

a6f1ed2

Update options.md

07116dc

Typo fix

Update options.md

a63c49c

Nicer formatting for options.

Update options.md

5b19748

Fixed unclosed formatting

Label clarifications

39e3a4b

Merge pull request #1978 from mikeboensel/patch-1

f71b3cf

Update options.md

Merge pull request #1979 from mikeboensel/patch-2

7a49955

Update options.md

Merge pull request #1980 from mikeboensel/patch-3

6a6c932

Update options.md

Merge pull request #1976 from jim60105/master

78e2df1

Optimize the Dockerfile to reduce the size of the image. (50.76 GB to 19.04 GB)

Merge pull request #1981 from mikeboensel/patch-4

2d0ed8e

Label clarifications

Merge branch 'dev' into DyLoRA-xl

86279c8

Merge pull request #1126 from tamlog06/DyLoRA-xl

488d187

Fix dylora create_modules error when training sdxl

fix to work with cpu_count() == 1 closes #1134

f413201

update einops to 0.7.0 #1122

24092e6

format by black

fb9110b

Merge branch 'dev' into resize_lora-add-rank-for-conv

0e70360

Merge pull request #1102 from mgz-dev/resize_lora-add-rank-for-conv

738c397

Resize lora add new rank for conv

fix format, add new conv rank to metadata comment

52b3799

some log output to print

8b7c142

fix ipex init

81e8af6

update readme

a21218b

Merge pull request #1136 from kohya-ss/dev

e69d341

v0.8.4

Merge branch 'main' of https://github.com/kohya-ss/sd-scripts into dev

a20c2bd

Merge branch 'dev' of https://github.com/bmaltais/kohya_ss into dev

822d94c

bmaltais merged commit 5d77bf4 into master Feb 24, 2024
0 of 2 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

v22.6.2 #1995

v22.6.2 #1995

bmaltais commented Feb 24, 2024

v22.6.2 #1995

v22.6.2 #1995

Conversation

bmaltais commented Feb 24, 2024