Stable DiffusionTips2

更新日2023-09-04 (月) 07:29:28
作成日:2023年8月16日

日本語化

日本語化.png

日本語化2.png

日本語化3.png

日本語化4.png

Beautiful Realistic Asiansモデルの追加

BRA5,BRA6を以下からDownload

BRA5

BRA6

ZIPを解凍して「models\Stable-diffusion\」に入れて再読み込み。

ControlnetのInstall

Controlnet1.png

Controlnet2.png

Controlnet3.png

Controlnet4.png

参考

https://koneko3.com/how-to-use-controlnet/

DynamicPromptのインストール

「Dynamic Prompt」はプロンプトの中に{|}を使うことで、1つの呪文(プロンプト)を何パターンかに分けて、それぞれを画像生成する

DynamicPromptInst1.png

DynamicPromptInst2.png

DynamicPromptInst3.png

DynamicPromptInst4.png

「Stable Diffusion web UI」のアップデート方法

参考

Controlnet モデル インストール

Controlnetをインストールしただけだとモデルが入っていないので、Downloadしてインストールする
14種類をDownloadする

ダウウンロード

以下のURLにアクセス

「controlnet」を検索する

model_inst1.png

以下のURLにジャンプする

https://huggingface.co/lllyasviel/ControlNet-v1-1

model_inst2.png

&color(red){stable-diffusion-webui-master\extensions\sd-webui-controlnet\modelsniha には「.yaml」ファイルが必要なのでDLのこと};

model_inst3.png

DLしたファイルをコピーする

model_inst4.png

動作確認

txt2imgでControlnetに適当な画像(プロンプトの内容に合うような画像)をドラッグ&ドロップして以下のようにしてみた

a businesswoman in the midst of a sales pitch, earnestly engaging and persuading her audience.

作成サンプル

cn_sample.png

メモリ不足対応

以下のエラー対応

RuntimeError: Not enough memory, use lower resolution (max approx. 512x512).
' Edit by JE2ISM
'set COMMANDLINE_ARGS=
set COMMANDLINE_ARGS=--medvram

参考

Error

A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.

webui-user.batに以下を追加

set COMMANDLINE_ARGS=--no-half

' Edit by JE2ISM
'set COMMANDLINE_ARGS=
set COMMANDLINE_ARGS=--medvram --no-half

今回は不要のようだが以下の変更も有効のようだ

Error1.png

メモリ不足対応2

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.00 MiB (GPU 0; 4.00 GiB total capacity; 17.73 GiB already allocated; 0 bytes free; 17.84 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

@echo off

'Edit by JE2ISM
set PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.6, max_split_size_mb:128 ←ここ

if not defined PYTHON (set PYTHON=python)
if not defined VENV_DIR (set "VENV_DIR=%~dp0%venv")

トップ   新規 一覧 検索 最終更新   ヘルプ   最終更新のRSS