Update app.py

app.py

@@ -3,18 +3,9 @@ from gpt import get_model
 
 
 model_small, tokenizer_small = get_model("gpt2")
-model_large, tokenizer_large = get_model("gpt2-large")
 
 def predict(inp, model_type):
-    if model_type == "gpt2-large":
-        model, tokenizer = model_large, tokenizer_large
-        input_ids = tokenizer.encode("user:"+inp+"\nai:", return_tensors='tf')
-        beam_output = model.generate(input_ids, max_length=60, num_beams=5,
-                                     no_repeat_ngram_size=2,
-                                     early_stopping=True)
-        output = tokenizer.decode(beam_output[0], skip_special_tokens=True,
-                                  clean_up_tokenization_spaces=True)
-    else:
+    if model_type == "gpt2":
         model, tokenizer = model_small, tokenizer_small
         input_ids = tokenizer.encode("user:"+inp+"\nai:", return_tensors='tf')
         beam_output = model.generate(input_ids, max_length=180, num_beams=5,
@@ -28,11 +19,11 @@ def predict(inp, model_type):
  
 
 INPUTS = [gradio.inputs.Textbox(lines=2, label="Input Text"),
-            gradio.inputs.Radio(choices=["gpt2-small", "gpt2-large"],
+            gradio.inputs.Radio(choices=["gpt2-small"],
                                 label="Model Size")]
 OUTPUTS = gradio.outputs.Textbox()
 examples = [
-    ["The toughest thing about software engineering is", "gpt2-large"],
+    ["The toughest thing about software engineering is", "gpt2-small"],
     ["Is this the real life? Is this just fantasy?", "gpt2-small"]
 ]
 INTERFACE = gradio.Interface(fn=predict, inputs=INPUTS, outputs=OUTPUTS, title="Chat GPT-2",