Fine-tune evaluation accuracy

Author: Anagha Jamthe

tutorials/Fine_Tuning_Vision/02-evaluation.md

@@ -1 +1,25 @@
 ## Evaluating the Fine-tuned Model 
+
+To try this part of the tutorial, we will use the fine-tuned model `best.pt` created by the ultralytics-fine-tune job.
+
+Your fine-tuned model will be accessible through Jupyter in this location
+`$WORK -> vista -> jobUUID -> train -> weights -> best.pt`
+
+You should already have the generated code that we tested accuracy on the base line model.
+
+Now, just change the path of your model to the fine-tuned model path and re-run the code to test the accuracy.
+
+When the code runs on the test data, you should see improved accuracy
+
+```
+Evaluation Metrics:
+Total images processed: 100
+Total animal images on (based label files): 71
+True Positives: 50
+True Negatives: 25
+False Positives: 4
+False Negatives: 21
+Overall detection accuracy: 0.75
+```
+
+As an exercise, you may train your model with more EPOCHS and see if this accuracy can be further improved.