ai-skills

Provides reference for evaluating AI-generated images and text-to-image (T2I) models. Covers mainstream benchmarks (PartiPrompt, GenAI-Bench, VQAScore, ImageEval, DrawBench, T2I-CompBench++, HRSBench, RAHF, BodyMetric, Human Distortion, Reason-Edit, etc.), metrics (FID, CLIPScore, VQAScore, human alignment, realism), and dataset scales. Use when the user asks about AI image evaluation, T2I benchmarks, image quality metrics, 图像评估, 文生图评测, benchmark, VQAScore, FID, 图文对齐, or how to evaluate generated images.

Generates Seedream 5.0/4.0 (即梦AI) image prompts and can call the API to generate images and auto-download to output/. Workflow: describe idea → agent outputs prompt for review → user confirms → agent runs generate.py. Covers text-to-image, image editing, multi-image fusion, character consistency, knowledge cards, posters, PPT backgrounds, e-commerce, avatars, group/storyboard. Use when the user mentions seedream, 即梦, jimeng, AI生图, 文生图, 图生图, seedream prompt, 提示词, 一键生图, knowledge card, poster design, 海报, 电商图, 角色一致性, or image generation.