There's two cross_fold inference results

[brifordwylie]

The pull_cv_results() will pull the CV S3 file from the training validation csv but I think the main way to pull the CV results is from model.get_inference_predictions()

wr.s3.read_csv("s3://sandbox-sageworks-artifacts/endpoints/open-admet-chemprop-logd/inference/full_cross_fold/inference_predictions.csv")
Out[13]:
     molecule_name  logd  prediction  prediction_std
0        E-0010647   1.5    1.401254        0.087501
1        E-0011026   2.8    3.120052        0.142324
2        E-0011083  -0.1   -0.169078        0.570826
3        E-0011133  -0.6   -0.773027        0.582125
4        E-0011137  -0.6   -0.257231        0.710500
...            ...   ...         ...             ...
5034     E-0020096   2.5    2.333825        0.113333
5035     E-0020097   1.5    1.645481        0.164296
5036     E-0020098   1.0    1.507061        0.201049
5037     E-0020099   2.2    2.049621        0.058354
5038     E-0020100   0.5    0.947165        0.207750