Update app.py

app.py

@@ -2,7 +2,7 @@ import torch
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 
-model_id = "openai-community/gpt2"  # можно flan-t5-large, если хватает памяти
+model_id = "google/flan-t5-base"  # можно flan-t5-large, если хватает памяти
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForSeq2SeqLM.from_pretrained(model_id)
 device = "cuda" if torch.cuda.is_available() else "cpu"
@@ -13,7 +13,7 @@ context = """
 специализирующийся на IT и робототехнике, расположенный в городе Иннополис, Татарстан.
 """
 
-def respond(message, history=None):
+def respond(message, history):
     if history is None:
         history = []
 
@@ -33,10 +33,18 @@ def respond(message, history=None):
             do_sample=False,
             eos_token_id=tokenizer.eos_token_id
         )
+
     answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
 
     history.append((message, answer))
-    return history
+    return history, history
+
+with gr.Blocks() as demo:
+    chatbot = gr.Chatbot()
+    message = gr.Textbox(placeholder="Введите вопрос...")
+    state = gr.State([])
+
+    message.submit(respond, inputs=[message, state], outputs=[chatbot, state])
+    message.submit(lambda: "", None, message)  # очистка поля ввода
 
-iface = gr.ChatInterface(fn=respond, title="Innopolis Q&A")
-iface.launch()
+demo.launch()