cleaning the input

app.py

@@ -1,3 +1,4 @@
+import re
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 
@@ -6,7 +7,8 @@ model = AutoModelForSequenceClassification.from_pretrained("usman-yello/job-tagg
 
 
 def tag_job_description(job_description):
-    inputs = tokenizer(job_description, return_tensors="pt")
+    job_description = re.sub(r'[^a-zA-Z]', ' ', job_description)
+    inputs = tokenizer(job_description, truncation=True, max_length=512, return_tensors="pt")
     outputs = model(**inputs)
     predicted_tag = model.config.id2label[outputs.logits.argmax().item()]
     return predicted_tag