Ethanfel f7ea559690 Speed: auto flash-attention/SDPA + document perf levers ...

transformers .generate() is the slow path; reasoning token volume and swap_eval
(2 passes) are the multipliers. Now requests attn_implementation flash_attention_2
-> sdpa -> default automatically (free speedup, flash-attn optional). README gains
a Performance section: swap_eval off (biggest free win), flash-attn, smaller model/
fewer axes, avoid nf4 for speed, and vLLM/SGLang as the real production-speed path.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

2026-06-27 11:18:11 +02:00

..

__init__.py

Initial commit: VLM-as-judge prompt calibration loop

2026-06-26 22:15:56 +02:00

qwen_judge.py

Speed: auto flash-attention/SDPA + document perf levers

2026-06-27 11:18:11 +02:00

receptor.py

Fix: render text inputs as editable fields, not input sockets

2026-06-27 10:18:12 +02:00