facebookresearch / facebookresearch/dlrm
[Question] Best ACC & AUC Never Get Updated in Inference State / Function?
- Dominant language
- Python
- Stars
- 4.1k
- Forks
- 859
- PR merge metrics
- No merged PRs in 30d
Description
https://github.com/facebookresearch/dlrm/blob/9c2fda79afbc09e277c17e420bffe510125b4f70/dlrm_s_pytorch.py#L1353-L1357
https://github.com/facebookresearch/dlrm/blob/9c2fda79afbc09e277c17e420bffe510125b4f70/dlrm_s_pytorch.py#L750-L759
https://github.com/facebookresearch/dlrm/blob/9c2fda79afbc09e277c17e420bffe510125b4f70/dlrm_s_pytorch.py#L880-L890
From the above code snippets, it looks like to me that the values **best_acc_test** and **best_auc_test** never get updated as it only stays in the inference function scope.
Is it designed for that? or if it is actually a bug?
Contributor guide
Research direction
Read dlrm_s_pytorch.py at lines 750-759, 880-890, and 1353-1357, then trace how best_acc_test and best_auc_test are created and used during inference. Done means determining whether their scope is intentional and documenting the conclusion or identifying the precise behavior that needs correction.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100