CalamitousFelicitousness 10771bd123 Merge branch 'dev' into feat/attention-core ...

dev added sdnq_attention_quantize_fp32 to modules/attention.py, which this
branch replaced with the modules/attention package. The option moves to the
sdnq backend in two places: the options dict the prepared call captures, and
the options tuple that rebuilds the chain when a captured option changes.
Reading it in only the first place would leave a setting that takes effect on
the next model load and not before.

The new option gets a hint alongside the rest of the section.

2026-08-24 22:54:59 +01:00

..

backends

Merge branch 'dev' into feat/attention-core

2026-08-24 22:54:59 +01:00

__init__.py

perf(attention): capture backend options once per chain

2026-08-23 08:21:28 +01:00

context.py

feat(attention): generation context

2026-08-23 08:20:46 +01:00

debug.py

feat(attention): route tracing and chain telemetry

2026-08-23 08:20:46 +01:00

dispatcher.py

chore(settings): remove the bmm attention methods

2026-08-24 19:59:41 +01:00

registry.py

perf(attention): capture backend options once per chain

2026-08-23 08:21:28 +01:00

router.py

perf(attention): capture backend options once per chain

2026-08-23 08:21:28 +01:00