alibaba / alibaba/FederatedScope
FS-LLM,拼写错误导致训练round数无法被正确赋值,怎么设置都只跑了10round
- Dominant language
- Python
- Stars
- 1.5k
- Forks
- 261
- PR merge metrics
- No merged PRs in 30d
Description
**Bug复现**
standalone模式下,多个GPU训练,线程数大于1, 会生成StandaloneMultiGPURunner
` if mode == 'standalone':
if process_num <= 1:
runner_cls = StandaloneRunner
else:
runner_cls = StandaloneMultiGPURunner`
然后再runner初始化的时候,
total_round_num 被写成了totol_round_num
server初始化这个参数是10。
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by locating StandaloneMultiGPURunner and inspect its initialization of the total round count, comparing it with the server's default of 10. Reproduce standalone multi-GPU training with more than one thread and verify that the configured round count is honored instead of always running 10 rounds.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- distributed-systems, machine-learning
- Issue type
- Bug
- Difficulty
- 1/5
- Estimated time
- Under an hour
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 58/100