Update models.py

apps/models.py

@@ -18,7 +18,7 @@ Additionally, in order to identify the root causes of the skill gap, curriculum
 
 ## Behavioral Analysis of the Model 
 
-We employed 15000 samples of data from 21 distinct types of job categories to train the model, which was was constructed via a transfer learning approach using the open-source **DistilBERT** transformer developed by researchers at Hugging Face.
+We employed 15000 samples of data from 21 distinct types of job categories to train the model, which was constructed via a transfer learning approach using the open-source **DistilBERT** transformer developed by researchers at Hugging Face.
 We used job requirements and other relevant data to train our final model. Resumes and curriculums were used to make gap predictions on the trained model. The percentage of matching between resumes and job requirements was shown to measure the gap in job supply and demand. All the skills were extracted using SkillNER based on the Spacy library.
 
 Model Limitation: One of the main limitations of the model is the dataset it was trained on. The original dataset had 62 categories, but due to insufficient data in many categories, some of them were combined, resulting in 21 categories. This approach of combining categories can make accurate CV segmentation more difficult. Additionally, the model was trained on an unbalanced dataset, which may lead to bias in certain situations. To overcome this limitation, larger and balanced datasets for each category would allow for more precise CV segmentation and lead to better output.