Model prompt change

app.py

@@ -27,7 +27,9 @@ tiny_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
 tokenizer = AutoTokenizer.from_pretrained(tiny_model)
 chat_model = AutoModelForCausalLM.from_pretrained(
     tiny_model,
-    torch_dtype=torch.float32)
+    torch_dtype=torch.bfloat16, 
+    device_map="auto",
+    low_cpu_mem_usage=True)
 
 pipe = pipeline(
     "text-generation",
@@ -73,6 +75,7 @@ def explain_recycling(class_label):
 
     outputs = pipe(
         prompt,
+        max_new_tokens=60,
         do_sample=True,
         temperature=0.3,
         top_p=0.9
@@ -192,7 +195,7 @@ with gr.Blocks() as demo:
         num_top_classes=3,
         label="🔍 Classifier Prediction (Top 3)"
     )
-    analyze_btn = gr.Button("Analyze Waste", variant="primary")
+    analyze_btn = gr.Button("Classify Waste", variant="primary")
 
     analyze_btn.click(
         classify_pipeline,