deepseek-ai/DeepEP
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepEP.git synced 2026-11-05 00:47:32 +00:00
Code Issues 308 Projects Packages Wiki Activity

[GH-ISSUE #615] official dockerfile #370

New issue
Open
opened 2026-10-26 09:59:17 +00:00 by bsakaguchi · 1 comment
bsakaguchi commented 2026-10-26 09:59:17 +00:00
Owner
Copy link

Originally created by @ybenvidia on GitHub (Apr 28, 2026).
Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/615

Hey,
Is it possible to have an official dockerfile based on pytorch with the minimal dependencies to build and run DeepEP ?

Originally created by @ybenvidia on GitHub (Apr 28, 2026). Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/615 Hey, Is it possible to have an official dockerfile based on pytorch with the minimal dependencies to build and run DeepEP ?
bsakaguchi commented 2026-10-26 09:59:31 +00:00
Author
Owner
Copy link

@YidaWeng commented on GitHub (Oct 1, 2026):

I would like to add a minimal official Dockerfile based on the PyTorch CUDA devel image.

Plan:

  • docker/Dockerfile on pytorch/pytorch:2.14.1-cuda13.2-cudnn9-devel (CUDA 13.2 toolkit for DeepJIT; PyTorch ≥ 2.10)
  • install nvidia-nccl-cu13>=2.32.3 --no-deps, then bash install.sh
  • short docker/README.md + a README pointer
  • document that the DeepJIT submodule must be initialized in the build context, and that RDMA/multi-node remains host-specific

I can develop and syntax-check this on macOS; a full image build/import smoke test still needs a Linux NVIDIA host. Is anyone already working on #615?

<!-- gh-comment-id:5929483990 --> @YidaWeng commented on GitHub (Oct 1, 2026): I would like to add a minimal official Dockerfile based on the PyTorch CUDA devel image. Plan: - `docker/Dockerfile` on `pytorch/pytorch:2.14.1-cuda13.2-cudnn9-devel` (CUDA 13.2 toolkit for DeepJIT; PyTorch ≥ 2.10) - install `nvidia-nccl-cu13>=2.32.3 --no-deps`, then `bash install.sh` - short `docker/README.md` + a README pointer - document that the DeepJIT submodule must be initialized in the build context, and that RDMA/multi-node remains host-specific I can develop and syntax-check this on macOS; a full image build/import smoke test still needs a Linux NVIDIA host. Is anyone already working on #615?
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
hybrid-ep
chore/remove-clang-format-yapf
v2.5
nvDev
try-fix-cu128
mori-ep
antgroup-opt
compat-old-torch
add-linter
eager-rdma
v1.2.1
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepEP#370
No description provided.