0z5a e58a728b80 feat: version rollout weight updates ...

Track a monotonic policy version across optimizer steps, scheduler updates, and rollout results. Serialize synchronous generation with weight acknowledgements and invalidate reusable prefix KV entries so cached samples remain attributable to the behavior policy that generated them.

2026-09-02 19:01:41 +08:00

..

conftest.py

test: refactor tests and fix inference edge cases

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

test_cache.py

feat: version rollout weight updates

2026-09-02 19:01:41 +08:00

test_engine.py

fix: cancel abandoned generation tasks

2026-09-02 12:55:05 +08:00

test_protocol.py

test: refactor tests and fix inference edge cases

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

test_sample.py

refactor: split infer core into subpackages by concern

2026-08-08 23:43:05 +08:00

test_scheduler.py

feat: version rollout weight updates

2026-09-02 19:01:41 +08:00

test_server.py

fix: cancel abandoned generation tasks

2026-09-02 12:55:05 +08:00

test_task.py

fix: cancel abandoned generation tasks

2026-09-02 12:55:05 +08:00

test_tool_parser.py

refactor: split infer core into subpackages by concern

2026-08-08 23:43:05 +08:00