Update README.md

README.md

@@ -41,11 +41,21 @@ has learned to predict binding sites well (and that EvoProtGrad works as intende
 
 ## Training
 
+This model was trained on approximately 70,000 proteins with binding site and active site annotations in UniProt. 
+The training split was a random 85/15 split for this version, and does not consider anything in the way of family or sequence 
+similarity. New iterations of the model have been trained on larger datasets (over 200,000 proteins), with the split such that 
+there are no overlapping families, however they seem to overfit much earlier and have significantly worse performance in terms 
+of the training metrics (precision, recall, and F1). 
+
+Training Metrics for the Model in the form of the `trainer_state.json` can be 
+[found here](https://huggingface.co/AmelieSchreiber/esm2_t6_8M_general_binding_sites_v2/blob/main/trainer_state.json).
+
 ```
 epoch 3:
 Training Loss Validation Loss Precision	Recall	 F1	      Auc
 0.031100	  0.074720	      0.684798	0.966856 0.801743 0.980853
 ```
+The hyperparameters are:
 
 ```
 wandb: 	lr: 0.0004977045729600779