Update app.py

app.py

@@ -39,7 +39,7 @@ def load_models():
                 FastLanguageModel.for_inference(model)
 
             except Exception as e:
-                st.error(f"⚠️ Failed to load Mistral model with Unsloth: {e}")
+                st.sidebar.error(f"⚠️ Failed to load Mistral model with Unsloth: {e}")
                 continue
         else:
             tokenizer = AutoTokenizer.from_pretrained(path)
@@ -49,8 +49,9 @@ def load_models():
         models[name] = {"tokenizer": tokenizer, "model": model}
     return models
 
-
+st.sidebar.write("📥 Load Models.")
 models = load_models()
+st.sidebar.write(f"✅ Successfully loaded models:{models}")
 
 model_choice = st.selectbox("Choose a model:", list(MODEL_OPTIONS.keys()))
 tokenizer = models[model_choice]["tokenizer"]
@@ -61,8 +62,9 @@ model = models[model_choice]["model"]
 def load_gsm8k_dataset():
     return load_dataset("openai/gsm8k", "main")["test"]
 
+st.sidebar.write("📥 Load GSM8K")
 gsm8k_data = load_gsm8k_dataset()
-st.write("📊 GSM8K loaded:", len(gsm8k_data), "samples")
+st.sidebar.write("📊 GSM8K loaded:", len(gsm8k_data), "samples")
 
 # === TABS ===
 tab1, tab2 = st.tabs(["🔓 Manual Prompting", "📊 GSM8K Evaluation"])
@@ -71,7 +73,7 @@ tab1, tab2 = st.tabs(["🔓 Manual Prompting", "📊 GSM8K Evaluation"])
 with tab1:
     prompt = st.text_area("Enter your math prompt:", "Jasper has 5 apples and eats 2 of them. How many apples does he have left?")
     if st.button("Generate Response", key="manual"):
-        with st.spinner("Generating..."):
+        with st.sidebar.spinner("🔄 Generating..."):
             inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
             output = model.generate(
                 **inputs,
@@ -98,23 +100,24 @@ with tab2:
 
     if st.button("Run GSM8K Sample"):
         try:
-            sample = random.choice(gsm8k_data)
-            question = sample["question"]
-            gold_answer = sample["answer"]
-
-            inputs = tokenizer(question, return_tensors="pt").to(model.device)
-
-            st.markdown(f"Create Output")
-            output = model.generate(
-                **inputs,
-                max_new_tokens=150,
-                temperature=0.7,
-                do_sample=True,
-                pad_token_id=tokenizer.eos_token_id,
-                eos_token_id=tokenizer.eos_token_id,
-            )
-            generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
-            response_only = generated_text[len(question):].strip()
+            with st.sidebar.spinner("🔄 Generating..."):
+                sample = random.choice(gsm8k_data)
+                question = sample["question"]
+                gold_answer = sample["answer"]
+    
+                inputs = tokenizer(question, return_tensors="pt").to(model.device)
+    
+                st.markdown(f"Create Output")
+                output = model.generate(
+                    **inputs,
+                    max_new_tokens=150,
+                    temperature=0.7,
+                    do_sample=True,
+                    pad_token_id=tokenizer.eos_token_id,
+                    eos_token_id=tokenizer.eos_token_id,
+                )
+                generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
+                response_only = generated_text[len(question):].strip()
 
             st.subheader("📌 GSM8K Question")
             st.markdown(question)