Suyog Gupta f94af0fb86 [AutoDeploy] Make all ranks agree on kv-cache size (#4007) ... * make all ranks agree on kv-cache size Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * lint Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * lint Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * lint Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * lint Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * minor cleanups Signed-off-by: Suyog Gupta <suyogg@nvidia.com> * use all_gather_object wrapper Signed-off-by: Suyog Gupta <suyogg@nvidia.com> --------- Signed-off-by: Suyog Gupta <suyogg@nvidia.com>		2025-05-02 04:07:28 +08:00
..
compile	feat: [AutoDeploy] generalizing cudagraph to multiple dynamic inputs (#3589)	2025-04-23 03:38:51 +08:00
custom_ops	feat:[AutoDeploy] Enhance RoPE support (#3115)	2025-04-11 23:51:24 +08:00
distributed	[AutoDeploy] Make all ranks agree on kv-cache size (#4007)	2025-05-02 04:07:28 +08:00
models	fix: [AutoDeploy] update hf loading for e_score_correction_bias (#3847)	2025-04-26 02:03:47 +08:00
shim	chore: move all distributed related codes into _torch.distributed directory (#3511)	2025-04-15 08:39:17 +08:00
transformations	[AutoDeploy] Make all ranks agree on kv-cache size (#4007)	2025-05-02 04:07:28 +08:00
utils	feat:[AutoDeploy] Enhance RoPE support (#3115)	2025-04-11 23:51:24 +08:00
__init__.py	Update TensorRT-LLM (#2820)	2025-02-25 21:21:49 +08:00