docs: Add version 0.12.0 to the change history

[kohya-ss]

Summary

Prepares the README change history for the 0.12.0 release.

• The "Changes planned for the next release" section of README / README-ja becomes the Version 0.12.0 (2026-09-24) entry. It contains everything since 0.11.1: OFTv2 / BOFT (OFTv2 / BOFT のネットワークモジュールを追加 #2357, OFTv2/BOFT follow-up: direct model_io import, docs, README #2427), per-subset timestep sampling offset (feat: add per-subset timestep_bias for semantic-aware timestep sampling #2401, Add --show_timesteps_offset and document offset behavior under shift/flux_shift #2410), Windows on ARM64 / RTX Spark support (Make opencv-python an optional dependency (Windows on ARM64 / RTX Spark support) #2430, Bump transformers/schedulefree/safetensors for Windows on ARM64 wheels, make dadaptation optional in tests #2431, Merge requirements-no-opencv.txt into requirements.txt with environment markers (Windows on ARM64 / RTX Spark support) #2433), the transformers 5 / diffusers 0.40 update (Update transformers to 5.5.4, diffusers to 0.40.0, accelerate to 1.15.0, huggingface-hub to 1.32.0 #2436), transformers 5.6+ support (Support transformers 5.6+ (CLIPTextModel layout wrapper, eager T5 attention) #2437), the sampler fixes (Fix the dpmsolver / dpmsingle samplers and give a clear error for lms without scipy #2438) and the removal of gen_img_diffusers.py (Remove gen_img_diffusers.py, the old SD1.x / SD2.x image generation script #2439).
• README-ja was missing the entry for Support transformers 5.6+ (CLIPTextModel layout wrapper, eager T5 attention) #2437 (transformers 5.6+ support) and still said that 5.6+ was not supported yet. The entry is added and the stale note removed, so the Japanese section now matches the English one (8 items each).

Please merge this before the dev → main PR for the release.

🤖 Generated with Claude Code