Switched template

app.py

@@ -1,158 +1,111 @@
 import gradio as gr
-import os
-import spaces
-from transformers import GemmaTokenizer, AutoModelForCausalLM
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
-from threading import Thread
-
-# Set an environment variable
-HF_TOKEN = os.environ.get("HF_TOKEN", None)
-
-
-DESCRIPTION = """
-<div>
-<h1 style="text-align: center;">Meta Llama3 8B</h1>
-<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct"><b>Meta Llama3 8b Chat</b></a>. Meta Llama3 is the new open LLM and comes in two sizes: 8b and 70b. Feel free to play with it, or duplicate to run privately!</p>
-<p>🔎 For more details about the Llama3 release and how to use the model with <code>transformers</code>, take a look <a href="https://huggingface.co/blog/llama3">at our blog post</a>.</p>
-<p>🦕 Looking for an even more powerful model? Check out the <a href="https://huggingface.co/chat/"><b>Hugging Chat</b></a> integration for Meta Llama 3 70b</p>
-</div>
-"""
-
-LICENSE = """
-<p/>
----
-Built with Meta Llama 3
-"""
-
-PLACEHOLDER = """
-<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
-   <img src="./images/val_speaking_transparent.gif" style="width: 80%; max-width: 550px; height: auto; opacity: 0.55;  "> 
-   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">Val</h1>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Hi i'm Val, ask me anything about working for VPS...</p>
-</div>
-"""
-
-
-css = """
-h1 {
-  text-align: center;
-  display: block;
-}
-#duplicate-button {
-  margin: auto;
-  color: white;
-  background: #1565c0;
-  border-radius: 100vh;
-}
-"""
-
-# Load the tokenizer and model
-tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct")
-model = AutoModelForCausalLM.from_pretrained(
-    "meta-llama/Meta-Llama-3-8B-Instruct", device_map="auto"
-)  # to("cuda:0")
-terminators = [tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<|eot_id|>")]
-
-
-@spaces.GPU(duration=120)
-def chat_llama3_8b(
-    message: str, history: list, temperature: float, max_new_tokens: int
-) -> str:
-    """
-    Generate a streaming response using the llama3-8b model.
-    Args:
-        message (str): The input message.
-        history (list): The conversation history used by ChatInterface.
-        temperature (float): The temperature for generating the response.
-        max_new_tokens (int): The maximum number of new tokens to generate.
-    Returns:
-        str: The generated response.
-    """
-    conversation = []
-    for user, assistant in 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").to(
-        model.device
+from http import HTTPStatus
+import dashscope
+from dashscope import Generation
+from dashscope.api_entities.dashscope_response import Role
+from typing import List, Optional, Tuple, Dict
+from urllib.error import HTTPError
+
+default_system = "You are a helpful assistant."
+
+YOUR_API_TOKEN = os.getenv("YOUR_API_TOKEN")
+dashscope.api_key = YOUR_API_TOKEN
+
+History = List[Tuple[str, str]]
+Messages = List[Dict[str, str]]
+
+
+def clear_session() -> History:
+    return "", []
+
+
+def modify_system_session(system: str) -> str:
+    if system is None or len(system) == 0:
+        system = default_system
+    return system, system, []
+
+
+def history_to_messages(history: History, system: str) -> Messages:
+    messages = [{"role": Role.SYSTEM, "content": system}]
+    for h in history:
+        messages.append({"role": Role.USER, "content": h[0]})
+        messages.append({"role": Role.ASSISTANT, "content": h[1]})
+    return messages
+
+
+def messages_to_history(messages: Messages) -> Tuple[str, History]:
+    assert messages[0]["role"] == Role.SYSTEM
+    system = messages[0]["content"]
+    history = []
+    for q, r in zip(messages[1::2], messages[2::2]):
+        history.append([q["content"], r["content"]])
+    return system, history
+
+
+def model_chat(
+    query: Optional[str], history: Optional[History], system: str
+) -> Tuple[str, str, History]:
+    if query is None:
+        query = ""
+    if history is None:
+        history = []
+    messages = history_to_messages(history, system)
+    messages.append({"role": Role.USER, "content": query})
+    gen = Generation.call(
+        model="qwen1.5-72b-chat",
+        messages=messages,
+        result_format="message",
+        stream=True,
     )
-
-    streamer = TextIteratorStreamer(
-        tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True
+    for response in gen:
+        if response.status_code == HTTPStatus.OK:
+            role = response.output.choices[0].message.role
+            response = response.output.choices[0].message.content
+            system, history = messages_to_history(
+                messages + [{"role": role, "content": response}]
+            )
+            yield "", history, system
+        else:
+            raise HTTPError(
+                "Request id: %s, Status code: %s, error code: %s, error message: %s"
+                % (
+                    response.request_id,
+                    response.status_code,
+                    response.code,
+                    response.message,
+                )
+            )
+
+
+with gr.Blocks() as demo:
+    gr.Markdown("""<center><font size=8>Qwen1.5-72B-Chat</center>""")
+    gr.Markdown(
+        """<center><font size=4>Qwen1.5-72B-Chat is the 72-billion parameter chat model of the Qwen series.</center>"""
     )
 
-    generate_kwargs = dict(
-        input_ids=input_ids,
-        streamer=streamer,
-        max_new_tokens=max_new_tokens,
-        do_sample=True,
-        temperature=temperature,
-        eos_token_id=terminators,
+    with gr.Row():
+        with gr.Column(scale=3):
+            system_input = gr.Textbox(value=default_system, lines=1, label="System")
+        with gr.Column(scale=1):
+            modify_system = gr.Button("🛠️ Set system prompt and clear history.", scale=2)
+        system_state = gr.Textbox(value=default_system, visible=False)
+    chatbot = gr.Chatbot(label="Qwen1.5-72B-Chat")
+    textbox = gr.Textbox(lines=2, label="Input")
+
+    with gr.Row():
+        clear_history = gr.Button("🧹 Clear history")
+        sumbit = gr.Button("🚀 Send")
+
+    sumbit.click(
+        model_chat,
+        inputs=[textbox, chatbot, system_state],
+        outputs=[textbox, chatbot, system_input],
     )
-    # This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash.
-    if temperature == 0:
-        generate_kwargs["do_sample"] = False
-
-    t = Thread(target=model.generate, kwargs=generate_kwargs)
-    t.start()
-
-    outputs = []
-    for text in streamer:
-        outputs.append(text)
-        # print(outputs)
-        yield "".join(outputs)
-
-
-# Gradio block
-chatbot = gr.Chatbot(height=450, placeholder=PLACEHOLDER, label="Gradio ChatInterface")
-
-with gr.Blocks(fill_height=True, css=css) as demo:
-
-    gr.Markdown(DESCRIPTION)
-    gr.DuplicateButton(
-        value="Duplicate Space for private use", elem_id="duplicate-button"
+    clear_history.click(fn=clear_session, inputs=[], outputs=[textbox, chatbot])
+    modify_system.click(
+        fn=modify_system_session,
+        inputs=[system_input],
+        outputs=[system_state, system_input, chatbot],
     )
-    gr.ChatInterface(
-        fn=chat_llama3_8b,
-        chatbot=chatbot,
-        fill_height=True,
-        additional_inputs_accordion=gr.Accordion(
-            label="⚙️ Parameters", open=False, render=False
-        ),
-        additional_inputs=[
-            gr.Slider(
-                minimum=0,
-                maximum=1,
-                step=0.1,
-                value=0.95,
-                label="Temperature",
-                render=False,
-            ),
-            gr.Slider(
-                minimum=128,
-                maximum=4096,
-                step=1,
-                value=512,
-                label="Max new tokens",
-                render=False,
-            ),
-        ],
-        examples=[
-            ["Where is the nearest ."],
-            ["Tell me about working for the Victorian Public Sector."],
-            ["How do I book leave?"],
-            ["Tell me about my organisations Disability Network"],
-            [""],
-        ],
-        cache_examples=False,
-    )
-
-    gr.Markdown(LICENSE)
 
-if __name__ == "__main__":
-    demo.launch()
+demo.queue(api_open=False).launch(max_threads=10, height=800, share=False)

requirements.txt

@@ -1,3 +1 @@
-accelerate
-transformers
-SentencePiece
+dashscope