Update README.md

README.md

@@ -10,10 +10,10 @@ pipeline_tag: text-classification
 **PPPSL-ESM2**(PPPSL, Prediction of prokaryotic protein subcellular localization) is a protein language model fine-tuned from [**ESM2**](https://github.com/facebookresearch/esm) pretrained model [(***facebook/esm2_t36_3B_UR50D***)](https://huggingface.co/facebook/esm2_t36_3B_UR50D) on a prokaryotic protein subcellular localization dataset.   
 
 **PPPSL-ESM2** achieved the following results:  
-Train Loss: 0.0015  
-Train Accuracy: 0.9893  
-Validation Loss: 0.0184  
-Validation Accuracy: 0.9702  
+Train Loss: 0.0148  
+Train Accuracy: 0.9923  
+Validation Loss: 0.0718  
+Validation Accuracy: 0.9893  
 Epoch: 20 
 # The dataset for training **PPPSL-ESM2**
 The full dataset contains 11,970 protein sequences, including Cellwall (87), Cytoplasmic (6,905), CYtoplasmic Membrane (2,567), Extracellular (1,085), Outer Membrane (758), and Periplasmic (568).