kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-13 22:18:36 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[None][chore] Revert "[None][fix] change allreduce workspace dtype to torch.int64 t… (#9538)

Browse Source 
Signed-off-by: Zhenhuan Chen <zhenhuanc@nvidia.com>
This commit is contained in:
Zhenhuan Chen 2025-11-28 16:45:23 +08:00 committed by GitHub
parent 0d3c0c2156
commit 7c3bb8534d
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tensorrt_llm/plugin/plugin.py
View File

@ -738,7 +738,7 @@ class CustomAllReduceHelper:
3 * lamport_buffers_size,
)
flag_buffer = torch.tensor([0, 0, 0, lamport_buffers_size, 0],
dtype=torch.int64,
dtype=torch.int,
device="cuda")
buffers = [ipc_buffers, ipc_barriers, lamport_buffers, flag_buffer]