hsliuustc0106 7db3a0cd5e [Performance] Add per-component DLO offload policy ...

Signed-off-by: hsliuustc0106 <liuhongsheng4@huawei.com>

2026-08-26 07:23:36 +08:00

..

__init__.py

[Model] Optimize MiniMax-H3 DLO component lifecycle (#6526)

2026-08-25 23:13:41 +08:00

base.py

[Performance] Add per-component DLO offload policy

2026-08-26 07:23:36 +08:00

block_discovery.py

[Feature] Distributed layerwise offload with DP multi-concurrency + mmap weight… (#5397)

2026-07-30 13:35:38 +00:00

cuda_host_registration.py

[Diffusion] Register HWR mmap for direct DLO H2D (#6591)

2026-08-25 23:30:11 +08:00

distributed_layerwise_backend.py

[Performance] Add per-component DLO offload policy

2026-08-26 07:23:36 +08:00

host_registration.py

[Diffusion] Register HWR mmap for direct DLO H2D (#6591)

2026-08-25 23:30:11 +08:00

layerwise_backend.py

[Diffusion][Quantization] Enable MiniMax-H3 global FP8 with DLO (#5910)

2026-08-15 23:01:26 +08:00

module_collector.py

feat(minimax-h3): enable 2x RTX 4090/5090 with DLO (#5764)

2026-08-06 09:02:29 +08:00

module_residency.py

[Model] Optimize MiniMax-H3 DLO component lifecycle (#6526)

2026-08-25 23:13:41 +08:00

offload_plan.py

[Diffusion] Add loader-owned host-weight plans for DLO (TP=1) (#6213)

2026-08-16 11:02:38 +08:00

sequential_backend.py

[Bugfix][Diffusion] Fix MiniMax-H3 model-level CPU offload residency (#6072)

2026-08-21 10:30:53 +00:00

startup.py

[Diffusion] Integrate no-AllGather DLO with Host Weight Runtime (#6486)

2026-08-25 08:03:53 +08:00

tensor_utils.py

[Feature] Distributed layerwise offload with DP multi-concurrency + mmap weight… (#5397)

2026-07-30 13:35:38 +00:00