modelscope / modelscope/ms-swift
multi-node gkd_training, 多节点gkd训练希望支持ray方法启动
Open
Nobody has claimed this yet.
enhancement
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
现状：多节点训练时候，一个节点做推理rollout，其他节点做训练，起任务的时候会比较难，需要挨个机器进去起脚本。
诉求：希望支持ray方法启动，只要在master节点里配置即可，减少操作成本。
以及vllm server启动支持ipv6吗，跨节点的时候ipv4的vllm_host_ip可以访问到，但是ipv6 unreachable,网络层没问题，应用层问题。
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by separating the requested Ray-based multi-node launch flow from the vLLM IPv6 connectivity question, then trace the existing training and server-start entry points. Done means master-node configuration can launch the distributed job without manual per-node scripts, with the IPv6 behavior and supported configuration documented or resolved.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- distributed-systems, machine-learning, networking
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100