Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -8,6 +8,21 @@ tags:
|
|
| 8 |
|
| 9 |
Features: morgan1024, physchem
|
| 10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
## Inference
|
| 12 |
|
| 13 |
```python
|
|
|
|
| 8 |
|
| 9 |
Features: morgan1024, physchem
|
| 10 |
|
| 11 |
+
Model: the original LazyPredict ExtraTrees model, a sklearn Pipeline
|
| 12 |
+
(ColumnTransformer + ExtraTreesRegressor, 100 trees). Unpickling
|
| 13 |
+
`model.pkl` requires `lazypredict` (listed in `requirements.txt`).
|
| 14 |
+
Inference uses `infer.py`, which builds a pandas DataFrame with
|
| 15 |
+
columns `feature_0`..`feature_1029` (Morgan 1024 first, then the 6
|
| 16 |
+
physchem descriptors in the order MolWt, MolLogP, TPSA, NumHDonors,
|
| 17 |
+
NumHAcceptors, NumRotatableBonds).
|
| 18 |
+
|
| 19 |
+
## Metrics (original LazyPredict run)
|
| 20 |
+
|
| 21 |
+
Test set (80/10/10 split, seed 42): RMSE 0.4449, MAE 0.1437, R2 0.8668.
|
| 22 |
+
FDA-12: RMSE 0.9637, MAE 0.6794 (11 of 12 compounds with a wetlab pIC50).
|
| 23 |
+
|
| 24 |
+
See `metrics.csv` and `fda.csv` in this repo for the full tables.
|
| 25 |
+
|
| 26 |
## Inference
|
| 27 |
|
| 28 |
```python
|