[BUG] <title>全参微调时的总batchsize是多少?
aoji0606 opened this issue · comments
aoji commented
是否已有关于该错误的issue或讨论? | Is there an existing issue / discussion for this?
- 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions
该问题是否在FAQ中有解答? | Is there an existing answer for this in FAQ?
- 我已经搜索过FAQ | I have searched FAQ
当前行为 | Current Behavior
No response
期望行为 | Expected Behavior
No response
复现方法 | Steps To Reproduce
No response
运行环境 | Environment
- OS:
- Python:
- Transformers:
- PyTorch:
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):
备注 | Anything else?
No response
aoji commented
我用自己的数据去ft,执行的finetune_ds.sh,设置的总bs=8*16,lr=1e-6,训完loss在0.8713,最后测试mme效果非常差,这是什么原因呢
qianyu chen commented
您的数据集和mme的差别有多大呢,您有测试过其他indomain数据集吗?我这里用单卡 batch为16,在refcoco上训练,最起码是有效果的。