Update app.py

app.py

@@ -1,64 +1,83 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForQuestionAnswering
+import torch
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+# -------------------------------
+# Модель суммаризации
+# -------------------------------
+sum_tokenizer = AutoTokenizer.from_pretrained("LaciaStudio/Lacia_sum_small_v1")
+sum_model = AutoModelForSeq2SeqLM.from_pretrained("LaciaStudio/Lacia_sum_small_v1")
 
+def summarize_document(file):
+    if file is None:
+        return "Файл не загружен."
+    # Открываем файл и читаем его содержимое
+    with open(file, "r", encoding="utf-8") as f:
+        text = f.read()
+    input_text = "summarize: " + text
+    inputs = sum_tokenizer(input_text, return_tensors="pt", max_length=512, truncation=True)
+    summary_ids = sum_model.generate(inputs["input_ids"], max_length=150, num_beams=4, early_stopping=True)
+    summary = sum_tokenizer.decode(summary_ids[0], skip_special_tokens=True)
+    return summary
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+# -------------------------------
+# Модель вопросов-ответов (Q&A)
+# -------------------------------
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+qa_tokenizer = AutoTokenizer.from_pretrained("LaciaStudio/Kaleidoscope_large_v1")
+qa_model = AutoModelForQuestionAnswering.from_pretrained("LaciaStudio/Kaleidoscope_large_v1")
+qa_model.to(device)
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+def answer_question(context, question):
+    inputs = qa_tokenizer(question, context, return_tensors="pt", truncation=True, max_length=384)
+    inputs = {k: v.to(device) for k, v in inputs.items()}
+    outputs = qa_model(**inputs)
+    start_index = torch.argmax(outputs.start_logits)
+    end_index = torch.argmax(outputs.end_logits)
+    answer_tokens = inputs["input_ids"][0][start_index:end_index + 1]
+    answer = qa_tokenizer.decode(answer_tokens, skip_special_tokens=True)
+    return answer
 
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+def answer_question_file(file, question):
+    if file is None:
+        return "Файл не загружен."
+    with open(file, "r", encoding="utf-8") as f:
+        context = f.read()
+    return answer_question(context, question)
 
+def answer_question_text(context, question):
+    return answer_question(context, question)
 
+# -------------------------------
+# Интерфейс Gradio
+# -------------------------------
+with gr.Blocks() as demo:
+    gr.Markdown("# Интерфейс для суммаризации и вопросов-ответов")
+    with gr.Row():
+        # Левая колонка – суммаризация
+        with gr.Column():
+            gr.Markdown("## Суммаризация документа")
+            file_input_sum = gr.File(label="Прикрепить файл для суммаризации", file_count="single", type="file")
+            summarize_button = gr.Button("Суммаризировать")
+            summary_output = gr.Textbox(label="Суммаризация", lines=10)
+            summarize_button.click(fn=summarize_document, inputs=file_input_sum, outputs=summary_output)
+        
+        # Правая колонка – Q&A с двумя вкладками
+        with gr.Column():
+            gr.Markdown("## Вопрос-ответ по документу")
+            with gr.Tabs():
+                with gr.Tab("Загрузить файл"):
+                    file_input_qa = gr.File(label="Прикрепить файл с документом", file_count="single", type="file")
+                    question_input_file = gr.Textbox(label="Введите вопрос", placeholder="Ваш вопрос здесь")
+                    answer_button_file = gr.Button("Получить ответ")
+                    answer_output_file = gr.Textbox(label="Ответ", lines=5)
+                    answer_button_file.click(fn=answer_question_file, inputs=[file_input_qa, question_input_file], outputs=answer_output_file)
+                with gr.Tab("Ввести текст"):
+                    context_input = gr.Textbox(label="Введите текст документа", lines=10, placeholder="Текст документа здесь")
+                    question_input_text = gr.Textbox(label="Введите вопрос", placeholder="Ваш вопрос здесь")
+                    answer_button_text = gr.Button("Получить ответ")
+                    answer_output_text = gr.Textbox(label="Ответ", lines=5)
+                    answer_button_text.click(fn=answer_question_text, inputs=[context_input, question_input_text], outputs=answer_output_text)
+                    
 if __name__ == "__main__":
     demo.launch()