deepmodeling / deepmodeling/Uni-Mol
运行unimol的finetune功能预测分子性质MAE有误
- Dominant language
- Python
- Stars
- 1.2k
- Forks
- 181
- PR merge metrics
- No merged PRs in 30d
Description
### Describe the bug
请教各位老师一个问题,我在运行unimol的finetune功能预测分子性质时,发现训练过程中日志文件输出的valid_mae,和使用相应的checkpoint.pt文件作为权重预测的MAE、MSE不同
比如在第2000 epoch,训练日志如下:
2025-04-02 08:39:15 | INFO | valid | epoch 2000 | valid on 'valid' subset | loss 0.242 | valid_mae 0.078 | valid_mse 0.024 | valid_agg_mae 0.078 | valid_agg_mse 0.024 | valid_agg_rmse 0.1564 | bsz 172 | num_updates 43876 | best_valid_agg_mae 0.071
但基于checkpoint2000.pt权重对valid数据集进行预测,得到的MAE为0.099,请问可能是什么原因?
### Uni-Mol Version
Uni-Mol2
### Expected behavior
-
### To Reproduce
_No response_
### Environment
_No response_
### Additional Context
_No response_
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.