baidu / baidu/information-extraction
完全按照readme操作，为什么loss降低到5就过拟合了
Open
- Dominant language
- Python
- Stars
- 330
- Forks
- 93
- PR merge metrics
- No merged PRs in 30d
Description
pass_num改到99999，不然100次就会停下来。
其余完全不变，使用GPU训练到400多个pass时，avg_cost达到最低的5.1，之后就越来越高，过拟合了。
请问这是为什么？不是说要训练到3以下吗？
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with the README's training instructions and locate the code that handles pass_num and reports avg_cost. Reproduce the GPU run described in the issue, then compare the loss curve around 400 passes with the README's expected value below 3; done means the cause of the divergence is identified and documented.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100