Update app.py

app.py

@@ -7,6 +7,8 @@ import torch
 import transformers
 transformers.utils.move_cache()
 device = "cuda" if torch.cuda.is_available() else "cpu"
+from huggingface_hub import HfFolder
+password1 = HfFolder.get_secret("password")
 
 if torch.cuda.is_available():
     torch.cuda.max_memory_allocated(device=device)
@@ -123,19 +125,40 @@ def search(term, num_results=1, lang="ko", advanced=True, sleep_interval=0, time
     return all_results
 
 
-client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
+client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+def respond1(
+    message,
+    history: list[tuple[str, str]],
+    system_message,
+    max_tokens,
+    temperature,
+    top_p,
+    password
+):
+    if password==password1:
+        messages = [{"role": "system", "content": "Your name is Chatchat.And your creator of you is Sung Yoon.In Korean, it is 정성윤.These are the instructions for you:"+system_message}]
 
-def model(message,history: list[tuple[str, str]],system_message,max_tokens,temperature,top_p):
-    
-    system_instructions1 = "<s>[SYSTEM]"+system_message+"And your name is Chatchat.Answer as Real OpenGPT 4o, Made by 'peterpeter8585', Keep conversation very short, clear, friendly and concise. The text provided is a request for a specific type of response from you, the virtual assistant. The request asks you to provide friendly responses. The expectation is that I will avoid introductions and start answering the query directly, elaborating on all aspects of the request. As an AI-powered assistant, your task is to generate responses that appear as if they are created by the user and not an AI-powered assistant.the instructions: [USER]"
-    """Performs a web search, feeds the results to a language model, and returns the answer."""
-    web_results = search(message)
-    web2 = ' '.join([f"Link: {res['link']}\nText: {res['text']}\n\n" for res in web_results])
-    formatted_prompt = system_instructions1 + message + "[WEB]" + str(web2) + "[OpenGPT 4o]"
-    stream = client.text_generation(formatted_prompt, max_new_tokens=max_tokens, stream=True, temperature=temperature, details=True, return_full_text=False, top_p=top_p)
-    token = "".join([response.token.text for response in stream if response.token.text != "</s>"])
-    yield token
-    
+        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})
+
+        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
 examples = [
     "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
     "An astronaut riding a green horse",
@@ -148,7 +171,37 @@ css="""
     max-width: 520px;
 }
 """
+def respond2(
+    message,
+    history: list[tuple[str, str]],
+    system_message,
+    max_tokens,
+    temperature,
+    top_p,
+):
+    messages = [{"role": "system", "content": "Your name is Chatchat.And, your made by SungYoon.In Korean, 정성윤.And these are the instructions.Whatever happens, you must follow it.:"+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})
+
+    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
 if torch.cuda.is_available():
     power_device = "GPU"
 else:
@@ -240,14 +293,13 @@ with gr.Blocks(css=css) as demo2:
         inputs = [prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
         outputs = [result]
     )
-BOT_AVATAR="OpenAI_Logo.png"
-People="user.png"
+
 
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
 """
 aa = gr.ChatInterface(
-    model,
+    respond1,
     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"),
@@ -259,11 +311,25 @@ aa = gr.ChatInterface(
             step=0.05,
             label="Top-p (nucleus sampling)",
         ),
+        gr.Textbox()
     ],
-    multimodal=True
 )
 
-
+ab= gr.ChatInterface(
+    respond2,
+    additional_inputs=[
+        gr.Textbox(value="You are a Programmer.You yave to only make programs that the user orders.Do not answer any other questions exept for questions about Python or other programming languages.Do not do any thing exept what I said.", label="System message",  interactive=False),
+        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)",
+        ),
+    ],
+)
 if __name__ == "__main__":
     with gr.Blocks() as ai:
         gr.TabbedInterface([aa, demo2], ["gpt4", "image create"])