Update app.py

app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-from transformers import AutoTokenizer, AutoModelForCausalLM
+from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
 import torch
 import spaces
 
@@ -14,31 +14,43 @@ model = AutoModelForCausalLM.from_pretrained(
 )
 
 @spaces.GPU
-def predict(message, history, max_tokens, temperature, top_p):
-    # Формирование чата из истории и нового сообщения
-    chat = [{"role": "user" if i % 2 == 0 else "assistant", "content": m} 
-            for i, (m, _) in enumerate(history)] + [{"role": "user", "content": message}]
-    
-    # Применение шаблона чата
-    prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
-    
-    # Кодирование входных данных
-    inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt").to(model.device)
-    
-    # Генерация ответа
-    outputs = model.generate(
-        input_ids=inputs, 
-        max_new_tokens=max_tokens,
+def generate(
+    message: str,
+    chat_history: list[tuple[str, str]],
+    max_new_tokens: int = 1024,
+    temperature: float = 0.6,
+    top_p: float = 0.9
+) -> Iterator[str]:
+    conversation = []
+    for user, assistant in chat_history:
+        conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
+    conversation.append({"role": "user", "content": message})
+
+    input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt")
+    if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
+        input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:]
+        gr.Warning(f"Trimmed input from conversation as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens.")
+    input_ids = input_ids.to(model.device)
+
+    streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
+    generate_kwargs = dict(
+        {"input_ids": input_ids},
+        streamer=streamer,
+        max_new_tokens=max_new_tokens,
         do_sample=True,
-        temperature=temperature,
         top_p=top_p,
+        temperature=temperature,
+        num_beams=1
     )
-    
-    # Декодирование результата
-    response = tokenizer.decode(outputs[0][len(inputs[0]):], skip_special_tokens=True)
-    
-    return response.strip().replace("assistant", "", 1)
+    t = Thread(target=model.generate, kwargs=generate_kwargs)
+    t.start()
+
+    outputs = []
+    for text in streamer:
+        outputs.append(text)
+        yield "".join(outputs)
 
+        
 # Настройка интерфейса Gradio
 iface = gr.ChatInterface(
     predict,