{"slug":"openlair-quantizing-models-bitsandbytes","source_name":"openlair/quantizing-models-bitsandbytes","name":"Openlair/Quantizing Models Bitsandbytes","description":"Quantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction with minimal accuracy loss. Use when GPU memory is limited, need to fit larger models, or want faster inference. Supports INT8, NF4, FP4 formats, QLoRA training, and 8-bit optimizers. Works with HuggingFace Transformers.","version":1,"lift":{"pass_rate_delta_pts":4.55,"pass_rate_pct":100,"total_cases":22,"passed_cases":22,"tokens_delta_pct":174.1,"turns_delta_pct":0,"verdict":"pass","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-07T19:39:50.413751+00:00"},"skill_score":1,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":4.55,"with_pass_pct":100,"without_pass_pct":95.5,"tokens_delta_pct":174.1,"turns_delta_pct":0,"total_cases":22,"cases_aggregated":22,"verdict":"pass","never_hurt":true,"completed_at":"2026-08-07T19:39:50.413751+00:00","run_id":"f045c738-2bb3-4f59-b49e-24d00831b2f3","version_number":1,"is_latest_version":true,"gate":null}],"trust":{"skill_safety":"passed","safety_status":"clean","intent_verdict":"safe","content_status":"clean","indexable":true},"license":"MIT","install_count":0,"manifest_hash":"406ccc4ecbf9d7f5552189ecc1724836cffb31e848ecc831a77990fb5d1d780c","raw_url":"https://app.decimal.ai/s/openlair-quantizing-models-bitsandbytes/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/openlair-quantizing-models-bitsandbytes"}