lilith/model-boss

History

autocommit 685a890cf8 docs(model-encyclopedia): 📝 Add benchmark results for Qwen3-VL-8B model to documentation Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>		2026-05-11 21:50:35 -07:00
..
benchmarks	docs(model-encyclopedia): 📝 Add benchmark results for Qwen3-VL-8B model to documentation	2026-05-11 21:50:35 -07:00
diffusion.md	docs(model-encyclopedia): 📝 Improve model encyclopedia documentation with expanded definitions, usage examples, and structured entries	2026-03-27 13:09:15 -07:00
embeddings.md	docs(model-encyclopedia): 📝 Improve model encyclopedia documentation with expanded definitions, usage examples, and structured entries	2026-03-27 13:09:15 -07:00
llms.md	docs(model-encyclopedia): 📝 Add benchmark metrics for qwen3.6-35b-a3b, qwen3.6-27b, and mistral-small-3.2-24b models to llms.md and their JSON files	2026-05-11 14:31:46 -07:00
README.md	docs(model-encyclopedia): 📝 Improve model encyclopedia documentation with expanded definitions, usage examples, and structured entries	2026-03-27 13:09:15 -07:00
vision.md	docs(model-encyclopedia): 📝 Add benchmark JSON files for qwen3-vl-8b-instruct and gemma-4-31b, and update vision.md with performance metrics	2026-05-11 14:31:46 -07:00

README.md

Model Encyclopedia

Operational notes about models managed by model-boss. Not API docs — this captures what we've learned running these models in production: quirks, VRAM gotchas, compatibility issues, and which models to use for what.

Index

Diffusion Models — SDXL, FLUX, SD3.5, ControlNet
Vision & Classification — SigLIP2, Qwen2.5-VL, Qwen3-VL, ImageReward
LLMs — Qwen3, DeepSeek, Nemotron, specialized models
Embeddings — Nomic Embed