upload custom README

README.md

@@ -1,7 +1,7 @@
 ---
+library_name: transformers
 base_model:
 - mercelisw/electra-grc
-library_name: transformers
 ---
 
 # Model Card for Model ID
@@ -13,13 +13,13 @@ This model is part of a series of models trained for the ML4AL paper “Gotta ca
 ### Model Description
 
 - **Developed by:** Marijke Beersmans & Alek Keersmaekers
-- **Model type:** ElectraForTokenClassification, finetuned for NER (GRP, PERS, LOC entities)
-- **Language(s) (NLP):** Ancient Greek (GLAUx normalization)
+- **Model type:** ElectraForTokenClassification, finetuned for NER (PERS, LOC, GRP).
+- **Language(s) (NLP):** Ancient Greek (greek_glaux normalization)
 - **Finetuned from model:** mercelisw/electra-grc
 
 ### Model Sources
 
-- **Repository:** [NERAncientGreekML4AL GitHub](https://github.com/NER-AncientLanguages/NERAncientGreekML4AL.git)(for data and training scripts)
+- **Repository:** [NERAncientGreekML4AL GitHub](https://github.com/NER-AncientLanguages/NERAncientGreekML4AL.git) (for data and training scripts)
 - **Paper:** [ML4AL paper](https://aclanthology.org/2024.ml4al-1.16/)
 
 ## Training Details
@@ -40,31 +40,27 @@ We thank the following projects for providing the training data:
 We use Weights & Biases for hyperparameter optimization with a random search strategy (10 folds), aiming to maximize the evaluation F1 score (eval_f1). 
 
 The search space includes:
-- Learning Rate: Sampled uniformly between 1e-6 and 1e-4
-- Weight Decay: One of [0.1, 0.01, 0.001]
-- Number of Training Epochs: One of [3, 4, 5, 6]
+ - Learning Rate: Sampled uniformly between 1e-6 and 1e-4
+ - Weight Decay: One of [0.1, 0.01, 0.001] 
+ - Number of Training Epochs: One of [3, 4, 5, 6]
 
 For the final training of this model, the hyperparameters were: 
-- Learning Rate: 9.889410158465026e-05
-- Weight Decay: 0.1
-- Number of Training Epochs: 5 
-
+ - Learning Rate: 9.889410158465026e-05
+ - Weight Decay: 0.1
+ - Number of Training Epochs: 5
 
 
 ## Evaluation
 
-This models was evaluation on precision, recall and macro-f1 for its entity classes. See the paper for more information.
-
-|              |   precision |   recall |   f1-score |   support |
+This models was evaluated on precision, recall and macro-f1 for its entity classes. See the paper for more information.
+| Label        |   precision |   recall |   f1-score |   support |
 |:-------------|------------:|---------:|-----------:|----------:|
-| GRP          |    0.778515 | 0.848266 |   0.811895 |      1384 |
-| LOC          |    0.708829 | 0.755656 |   0.731494 |      1105 |
-| PERS         |    0.845869 | 0.888026 |   0.866435 |      3090 |
-| micro avg    |    0.801518 | 0.851945 |   0.825962 |      5579 |
-| macro avg    |    0.777737 | 0.830649 |   0.803275 |      5579 |
-| weighted avg |    0.802018 | 0.851945 |   0.826178 |      5579 |
-
-
+| GRP          |      0.8054 |   0.8013 |     0.8033 |      1384 |
+| LOC          |      0.7379 |   0.6905 |     0.7134 |      1105 |
+| PERS         |      0.853  |   0.866  |     0.8595 |      3090 |
+| micro avg    |      0.8198 |   0.8152 |     0.8175 |      5579 |
+| macro avg    |      0.7988 |   0.7859 |     0.7921 |      5579 |
+| weighted avg |      0.8184 |   0.8152 |     0.8166 |      5579 |
 
 
 If you use this work, please cite the following paper:
@@ -87,5 +83,4 @@ Beersmans, M., Keersmaekers, A., de Graaf, E., Van de Cruys, T., Depauw, M., & F
   year = {2024},
   month = aug,
   pages = {152--164}
-}
-
+}