Update app.py

app.py

@@ -3,57 +3,14 @@ from transformers import GPT2LMHeadModel, GPT2Tokenizer
 import torch
 
 # Initialisierung des Modells und des Tokenizers
-tokenizer = GPT2Tokenizer.from_pretrained("Loewolf/GPT1")
-model = GPT2LMHeadModel.from_pretrained("Loewolf/GPT1")
-model.to("cpu")  # Stellen Sie sicher, dass das Modell auf der CPU läuft
+tokenizer = GPT2Tokenizer.from_pretrained("gpt2")
+model = GPT2LMHeadModel.from_pretrained("gpt2")
 
-# Chat-Verlauf initialisieren
-chat_history = []
-
-def generate_text(input_text, history, temperature, top_k, top_p, max_length):
-    global chat_history
-    # Hinzufügen der neuen Eingabe zum Chat-Verlauf
-    chat_history.append(f"Nutzer: {input_text}")
-    new_input = tokenizer.encode(input_text + tokenizer.eos_token, return_tensors='pt')
-    new_input = new_input.to("cpu")  # Stellen Sie sicher, dass Tensoren auf der CPU sind
-
-    # Generieren der Antwort des Modells
-    chat_history_ids = tokenizer.encode(" ".join(chat_history) + tokenizer.eos_token, return_tensors='pt')
-    chat_history_ids = chat_history_ids.to("cpu")
-    
-    # Achten Sie auf die Größe der Tokens für das Modell
-    chat_history_ids = chat_history_ids[:, -tokenizer.model_max_length :]
-    
-    # Generieren der Antwort
-    model_output = model.generate(chat_history_ids, max_length=max_length, pad_token_id=tokenizer.eos_token_id, 
-                                  temperature=temperature, top_k=top_k, top_p=top_p)
-    
-    # Antwort in Text umwandeln
-    response = tokenizer.decode(model_output[:, chat_history_ids.shape[-1]:][0], skip_special_tokens=True)
-    chat_history.append(f"Löwolf GPT: {response}")
-    
-    # Rückgabe des aktualisierten Chat-Verlaufs
-    return " ".join(chat_history)
-
-# Erstellen der Gradio-Schnittstelle
-with gr.Blocks() as demo:
-    with gr.Row():
-        with gr.Column():
-            history = gr.Textbox(label="Chatverlauf", value=" ".join(chat_history), lines=10, interactive=False)
-            user_input = gr.Textbox(label="Deine Nachricht")
-            submit_btn = gr.Button("Senden")
-        with gr.Column():
-            temperature = gr.Slider(minimum=0, maximum=1, step=0.01, label="Temperature", value=0.7)
-            top_k = gr.Slider(minimum=0, maximum=100, step=1, label="Top K", value=50)
-            top_p = gr.Slider(minimum=0, maximum=1, step=0.01, label="Top P", value=0.9)
-            max_length = gr.Slider(minimum=1, maximum=100, step=1, label="Maximale Länge", value=60)
-
-    submit_btn.click(
-        generate_text,
-        inputs=[user_input, history, temperature, top_k, top_p, max_length],
-        outputs=[history]
-    )
-
-# Starten der Gradio-App
-demo.launch()
+def generate_text(prompt):
+    inputs = tokenizer.encode(prompt, return_tensors="pt")
+    outputs = model.generate(inputs, max_length=50, num_return_sequences=1)
+    text = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return text
 
+iface = gr.Interface(fn=generate_text, inputs="text", outputs="text")
+iface.launch()