dharak-cohereandCursor d0e695a91b [Distributed] Support pre-shared ncclUniqueId rendezvous for weight transfer (#53784) ...

Signed-off-by: Dharak Kharod <dharak.kharod@cohere.com>
Co-authored-by: Cursor <cursoragent@cursor.com>

2026-09-01 14:55:57 +08:00

..

__init__.py

[rl] Stateful Trainer Send: New Abstractions [1/N] (#48042)

2026-07-17 15:11:06 +08:00

base.py

[RL] Add rank-local IPC weight updates (#52497)

2026-08-26 15:36:57 +08:00

clients.py

[RL] Add rank-local IPC weight updates (#52497)

2026-08-26 15:36:57 +08:00

factory.py

[RL] P2P RDT weight sync (#43375)

2026-08-23 09:42:22 -07:00

ipc_engine.py

[Docs][RL] Rewrite weight-transfer docs; standardize examples (#51729)

2026-08-11 18:53:15 -07:00

nccl_common.py

[Distributed] Support pre-shared ncclUniqueId rendezvous for weight transfer (#53784)

2026-09-01 14:55:57 +08:00

nccl_engine.py

[Distributed] Support pre-shared ncclUniqueId rendezvous for weight transfer (#53784)

2026-09-01 14:55:57 +08:00

packed_tensor.py

[Bugfix] Reuse CUDA streams in packed weight transfer to cap reserved-memory waste (#52951)

2026-08-23 23:54:41 -07:00

sharded_rdt_common.py

[RL] P2P RDT weight sync (#43375)

2026-08-23 09:42:22 -07:00

sharded_rdt_engine.py

[RL] P2P RDT weight sync (#43375)

2026-08-23 09:42:22 -07:00

sharded_rdt_fake.py

[RL] P2P RDT weight sync (#43375)

2026-08-23 09:42:22 -07:00

sharded_rdt_trainer.py

[RL] P2P RDT weight sync (#43375)

2026-08-23 09:42:22 -07:00

sparse_nccl_engine.py

[Distributed] Support pre-shared ncclUniqueId rendezvous for weight transfer (#53784)

2026-09-01 14:55:57 +08:00