Chapter 97: Uncategorized contents

Usually "not article".

Article for situational use

• A SoK on vid2vid(and seperate from text2vid): vid2vid.md

• A list shouldn't exist: 0299d.md

• A flowchart about current "APP" making use of "AI art" definitely not face swap : swap_face_app.md

• A quiz, meanwhile introduction to "SD in AI / ML": quiz_4751.md

• A block diagram about "RAG with Docuement versioning control": rag_with_doc.md

Scripts (Convert to diffuser and upload to HuggingFace later)

• convert_sdxl_to_diffusers.py: My fork of convert_sdxl_to_diffusers.py. Added quick guide on the top.
• convert_original_stable_diffusion_to_diffusers.py: My fork of convert_original_stable_diffusion_to_diffusers.py. Added quick guide on the top.

Scripts (surviving in the internet)

• purgeme.py: "/purgeme". Related to trojblue/telegram-scraper but not scraping.
• PixivDataAnalysis.py: Statistics without Pixiv Premium. Prototype of trojblue/pixivAnalytics.

Scripts (manual scripts for webui)

• step.js [step] [split_comma]: I use it frequently. Generate global static emphasis level on the prompts.
• node listseed [full_directory]: Extract seeds inside the filenames, which reduce human error when iterlating mass image production. Use [seed]-[width]-[height]-[cfg]-[steps]-[datetime] as Settings > Images filename pattern.
• aspect512.js [w] [h]: Find actual width / height with aspect ratio w:h under 512x512 px also in unit of 64px.

Scripts (broken shell scripts)

• anifusion2_win11_vscode_miniconda.sh: Install Anifusion-SD with Win11 + conda. Run it line by line in manual.

Scripts (reading metadata for a model by batch)

• extract_merge_info.js: Extract the actual JSON from safetensors_util. Author closed my PR. Fine.
• batch_extract.cmd: By batch. May fail a lot.