Update README.md

README.md

@@ -17,14 +17,14 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on the rishitunu/ecc_crackdetector dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1935
-- Mean Iou: 0.1871
-- Mean Accuracy: 0.3741
-- Overall Accuracy: 0.3741
+- Loss: 0.0351
+- Mean Iou: 0.9171
+- Mean Accuracy: 0.8041
+- Overall Accuracy: 0.8041
 - Accuracy Background: nan
-- Accuracy Crack: 0.3741
+- Accuracy Crack: 0.8041
 - Iou Background: 0.0
-- Iou Crack: 0.3741
+- Iou Crack: 0.9171
 
 ## Model description
 
@@ -45,7 +45,7 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 6e-05
 - train_batch_size: 2
-- eval_batch_size: 1
+- eval_batch_size: 2
 - seed: 1337
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: polynomial