Update app.py

app.py

@@ -1,60 +1,60 @@
 import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer
-import torch
-
-# モデルとトークナイザーの読み込み
-model_name = "Qwen/Qwen2.5-0.5b-Instruct"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name, ignore_mismatched_sizes=True)
-
-# 応答を生成する関数
-def respond(message, history, max_tokens, temperature, top_p):
-    # 入力履歴と新しいメッセージを連結
-    if history is None:
-        history = []
-
-    input_text = ""
-    for user_message, bot_response in history:
-        input_text += f"User: {user_message}\nAssistant: {bot_response}\n"
-    input_text += f"User: {message}\nAssistant:"
-
-    # トークナイズ
-    inputs = tokenizer(input_text, return_tensors="pt")
-
-    # モデルによる応答生成
-    with torch.no_grad():
-        outputs = model.generate(
-            inputs.input_ids,
-            max_length=inputs.input_ids.shape[1] + max_tokens,
-            do_sample=True,
-            top_p=top_p,
-            temperature=temperature,
-        )
-
-    # 応答をデコード
-    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    # 最後のユーザー入力以降の応答部分を抽出
-    response = response.split("Assistant:")[-1].strip()
-
-    # 応答と履歴を更新
-    history.append((message, response))
-    return response, history
-
-# Gradioインターフェースの設定
-with gr.Blocks() as demo:
-    gr.Markdown("## AIチャット")
-    chatbot = gr.Chatbot()
-    msg = gr.Textbox(label="あなたのメッセージ", placeholder="ここにメッセージを入力...")
-    max_tokens = gr.Slider(1, 2048, value=512, step=1, label="Max new tokens")
-    temperature = gr.Slider(0.1, 4.0, value=0.7, step=0.1, label="Temperature")
-    top_p = gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
-    send_button = gr.Button("送信")
-    clear = gr.Button("クリア")
-
-    def clear_history():
-        return [], []
-
-    send_button.click(respond, inputs=[msg, chatbot, max_tokens, temperature, top_p], outputs=[chatbot, chatbot])
-    clear.click(clear_history, outputs=[chatbot])
-
-demo.launch()
+from huggingface_hub import InferenceClient
+
+"""
+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("Qwen/Qwen2.5-0.5b-Instruct")
+
+
+def respond(
+    message,
+    history: list[tuple[str, str]],
+    system_message,
+    max_tokens,
+    temperature,
+    top_p,
+):
+    messages = [{"role": "system", "content": system_message}]
+
+    for val in history:
+        if val[0]:
+            messages.append({"role": "user", "content": val[0]})
+        if val[1]:
+            messages.append({"role": "assistant", "content": val[1]})
+
+    messages.append({"role": "user", "content": message})
+
+    # ストリーミングを無効にして、単一の応答を取得
+    response = client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        temperature=temperature,
+        top_p=top_p,
+    )
+
+    return response.choices[0].message.content
+
+"""
+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="ユーザーの質問や依頼にのみ答えてください。ポジティブに答えてください", label="System message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="新規トークン最大"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="温度"),
+        gr.Slider(
+            minimum=0.1,
+            maximum=1.0,
+            value=0.95,
+            step=0.05,
+            label="Top-p (核 sampling)",
+        ),
+    ],
+    concurrency_limit=24  # 例: 同時に4つのリクエストを処理
+)
+
+
+if __name__ == "__main__":
+    demo.launch()