Update app.py

app.py

@@ -27,28 +27,28 @@ modelconf = ModelConfig(**model_args)
 trained_model = BigramLanguageModel(modelconf)
 trained_model.load_state_dict(model_weights)
 
-def generate_text(seed_text, max_new_tokens, temperature):
+def generate_text(seed_text, max_new_tokens, confidence):
     text = seed_text if seed_text is not None else " "
     text = text if text.endswith(" ") else seed_text + " "
     context = torch.tensor(encode(text), dtype=torch.long).unsqueeze(0)
-    temperature = temperature if temperature > 0 else 1e-5
-    return decode(trained_model.generate(context, temperature = temperature, max_new_tokens=max_new_tokens)[0].tolist())
+    confidence = confidence if confidence > 0 else 1e-5
+    return decode(trained_model.generate(context, confidence = confidence, max_new_tokens=max_new_tokens)[0].tolist())
 
 with gr.Blocks() as demo:
-    gr.HTML("<h1 align = 'center'> Generate Text Based on simple GPT model <br> (Dataset = William Shakespeare) </h1>")
+    gr.HTML("<h1 align = 'center'> Simple GPT from scratch using tiny Shakespere </h1>")
     
-    content = gr.Textbox(label = "Enter initial text to generate content")
+    content = gr.Textbox(label = "Initial text to generate content")
     with gr.Row(equal_height=True):
       with gr.Column(): 
-        max_tokens = gr.Number(label = "Maximum tokens to generate content", value = 100)
-        temp_val = gr.Slider(label = "Temparature (slide to higher value for higher creativity)", minimum = 0.0, maximum= 1.0,value = 0.7)
+        max_tokens = gr.Number(label = "Maximum number of tokens", value = 100)
+        confidence = gr.Slider(label = "Confidence", minimum = 0.0, maximum= 1.0,value = 0.7)
         generate_btn = gr.Button(value = 'Generate Text')
       with gr.Column(): 
-        outputs  = [gr.TextArea(label = "Generated text (William Shakespeare)", lines = 7)]
+        outputs  = [gr.TextArea(label = "Generated result", lines = 8)]
     inputs = [
                 content,
                 max_tokens, 
-                temp_val
+                confidence
               ]
     generate_btn.click(fn = generate_text, inputs= inputs, outputs = outputs)