ViperEkura ae7fc3059a refactor: harden inference cache state and attention dispatch ...

- split KVCache into phase-specific PrefillKVCache/DecodeKVCache types selected by start_pos
- unify steady-state detection in TaskCacheManager
- guard decode steady-state reuse with the cached task signature so recycled req slots cannot replay a prior generation's tokens and positions
- collapse attention backend fwd_decode/fwd_prefill into a single subclass-owned forward with a shared _check_fwd guard
- fix thread-safety gap in weight update and validate prefill inputs before KV allocation
- centralize magic constants in InferenceConfig and align docs with behavior

2026-09-04 14:28:04 +08:00

..

conftest.py

test: refactor tests and fix inference edge cases

2026-08-21 22:59:51 +08:00

test_cache.py

test: deduplicate suites and prune low-value cases

2026-09-03 21:54:14 +08:00

test_engine.py

refactor: assemble inference engines through a shared composition root

2026-09-03 22:16:56 +08:00

test_protocol.py

test: deduplicate suites and prune low-value cases

2026-09-03 21:54:14 +08:00

test_sample.py

test: deduplicate suites and prune low-value cases

2026-09-03 21:54:14 +08:00

test_scheduler.py

refactor: harden inference cache state and attention dispatch

2026-09-04 14:28:04 +08:00

test_server.py

refactor: assemble inference engines through a shared composition root

2026-09-03 22:16:56 +08:00

test_task.py

fix: resolve audited training and inference bugs

2026-09-02 21:25:01 +08:00

test_tool_parser.py

test: deduplicate suites and prune low-value cases

2026-09-03 21:54:14 +08:00