Update app.py

app.py

@@ -1,34 +1,59 @@
 import os
-from transformers import pipeline
+import torch
 import gradio as gr
 
-HF_TOKEN = os.getenv("HF_TOKEN")
-
-qa = pipeline(
-    "question-answering",
-    model="harishforaiandml/my-pretrained-qa-model",
-    token=HF_TOKEN   # 👈 IMPORTANT FIX
+from transformers import (
+    AutoTokenizer,
+    AutoModelForQuestionAnswering
 )
 
+MODEL_NAME = "harishforaiandml/my-pretrained-qa-model"
+
+# Load tokenizer + model manually (SAFE for Spaces)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForQuestionAnswering.from_pretrained(MODEL_NAME)
+
+model.eval()
+
+
 def ask(context, question):
 
-    result = qa(
-        question=question,
-        context=context
+    inputs = tokenizer(
+        question,
+        context,
+        return_tensors="pt",
+        truncation=True
     )
 
-    if result["score"] < 0.1:
+    with torch.no_grad():
+        outputs = model(**inputs)
+
+    start_logits = outputs.start_logits
+    end_logits = outputs.end_logits
+
+    start_idx = torch.argmax(start_logits)
+    end_idx = torch.argmax(end_logits) + 1
+
+    answer_tokens = inputs["input_ids"][0][start_idx:end_idx]
+    answer = tokenizer.decode(answer_tokens, skip_special_tokens=True)
+
+    if answer.strip() == "":
         return "Not a valid question"
 
-    return result["answer"]
+    return answer
+
 
 demo = gr.Interface(
     fn=ask,
+
     inputs=[
         gr.Textbox(lines=6, label="Context"),
         gr.Textbox(label="Question")
     ],
+
     outputs="text",
+
+    title="QA Model (Your HF Model)"
 )
 
 demo.launch()