Replace cpu_offload constructor param with turn_on/turn_off API (#26)

draft commit for cpu_offload (#23)

Replace toy PP tests with real-model-based pipeline tests [skip-build]

Add correctness verification to PP tests using fully_shard [skip-build]

Remove correctness check from PP tests, focus on deadlock detection [skip-build]

Add PP + dp_replicate deadlock regression tests [skip-build]

Apply pre-commit formatting (isort) [skip-build]

Add MoE uneven shard test with mixed expert and non-expert params [skip-build]

Add uneven shard correctness test [skip-build]

Update tests for MoE and parallel optimizations [skip-build]

Add torch.compile, CUDA graph, and compiled momentum [skip-build]

Apply suggestions from code review

Support multi-component expert_keys (e.g. "experts.w1")

Use component-level matching for expert_keys to avoid shared_experts collision

Normalize parameter FQNs to handle torch.compile / checkpoint wrappers

Refactor pipeline to async generator pattern (#16)

Support mHC (#15)

Support param group with various placements (#13)

misc

apply all2all scatter gather

feat(muon_clip) : add muon clip (#6)

feat: update muon to receive paramgroups, not model (#4)

feat(muon): add test for muon