Skip to content

Commit

Permalink
Update model documentation
Browse files Browse the repository at this point in the history
  • Loading branch information
JulienPeloton committed Dec 3, 2024
1 parent 7b534f3 commit be1f4be
Showing 1 changed file with 8 additions and 8 deletions.
16 changes: 8 additions & 8 deletions fink_science/data/models/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,23 +2,23 @@

To load and run model properly, you need to check carefuly the dependencies used in Fink: [deps](https://github.com/astrolabsoftware/fink-broker/tree/master/deps).

Information about the science modules (and related journal papers), can be found in the [Fink documentation](https://fink-broker.readthedocs.io/en/latest/broker/science_modules/).

## ZTF

Default models used for ZTF:

| model name | used in | Description |
|------------|---------|-------------|
| `snn_models/snn_snia_vs_nonia/model.pt`| `snn` | Ia vs core-collapse SNe |
| `snn_models/snn_sn_vs_all/model.pt`| `snn` | SNe vs. anything else (var star and other stuff in training) |
| `default-model_sigmoid.obj` | `random_forest_snia` | Early SN Ia classification using lightcurve features (random forest). Binary classification. `sklearn==1.0.2` |
| `rf.sav` | `microlensing` | scikit-learn model used in Random Forest. `sklearn==1.0.2` |
| `pca.sav` | `microlensing` | scikit-learn (pca) model used in Random Forest. `sklearn==1.0.2` |
|`components.csv`| `kilonova`| file containing principal components from simulations|
| `snn_models/snn_snia_vs_nonia/model.pt`| `snn` | Binary classifier based on SuperNNova. Ia vs core-collapse SNe |
| `snn_models/snn_sn_vs_all/model.pt`| `snn` | Binary classifier based on SuperNNova. SNe vs. anything else (var star and other stuff in training) |
| `default-model_sigmoid.obj` | `random_forest_snia` | Early SN Ia classification using lightcurve features. Binary classification. Random forest. `sklearn==1.0.2` |
| `rf.sav` | `microlensing` | Random Forest classifier for microlensing classification. 4 classes: ML, VAR, CONST, SN. `sklearn==1.0.2` |
| `pca.sav` | `microlensing` | PCA for microlensing classification. `sklearn==1.0.2` |
| `for_al_loop/*.pkl` | active learning loop | Models used in the context of Active Learning for early SN Ia detection. `sklearn==1.0.2` |
|`anomaly_detection`| `anomaly_detection` | Zip files with different models used for Anomaly detection |
|`anomaly_detection`| `anomaly_detection` | Zip files with different models used for Anomaly detection. Based on modified Random Forest. |

Models for Kilonova are in [kndetect](https://github.com/b-biswas/kndetect).
Models for Kilonova detection are in [kndetect](https://github.com/b-biswas/kndetect).

## Elasticc/Rubin

Expand Down

0 comments on commit be1f4be

Please sign in to comment.