Update app.py

app.py

@@ -1,36 +1,143 @@
+from threading import Thread
+from typing import Iterator
+
 import gradio as gr
-from huggingface_hub import InferenceClient
+import spaces
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
 
-client = InferenceClient(model="wangzhang/chatSDB-test")
+MAX_MAX_NEW_TOKENS = 2048
+DEFAULT_MAX_NEW_TOKENS = 1024
+MAX_INPUT_TOKEN_LENGTH = 4096
+
+DESCRIPTION = """\
+# ChatSDB
+
+这是SequioaDB旗下的AI智能大语言模型，训练超过上万条真实数据和7亿参数。
+ChatSDB是SequoiaDB旗下的AI智能大语言模型，训练超过上万条真实数据和7亿参数</h3>
+<br><strong>模型🔗: <a>https://huggingface.co/wangzhang/ChatSDB </a></strong>
+<br><strong>Dataset🔗: <a>https://huggingface.co/datasets/wangzhang/sdb </a></strong>
+<br><strong> API Doc🔗: <a>https://zgg3nzdpswxy4a-80.proxy.runpod.net/docs/ <a> </strong>
+"""
+
+LICENSE = """ """
+
+if not torch.cuda.is_available():
+    DESCRIPTION += "\n<p>Running on CPU 🥶 This demo does not work on CPU.</p>"
+
+
+if torch.cuda.is_available():
+    model_id = "wangzhang/ChatSDB-hf"
+    model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto")
+    tokenizer = AutoTokenizer.from_pretrained(model_id)
+    tokenizer.use_default_system_prompt = False
 
-def inference(message, history):
-    prompt = f"""### Instruction:
 
-            ### Task:
-            根据巨杉数据库SequoiaDB的相关问题进行回答。
-            
-            ### Input:
-            {message}
-            
-            ### Response:
+@spaces.GPU
+def generate(
+    message: str,
+    chat_history: list[tuple[str, str]],
+    system_prompt: str,
+    max_new_tokens: int = 512,
+    temperature: float = 0.2,
+    top_p: float = 0.9,
+    top_k: int = 10,
+    repetition_penalty: float = 1.2,
+) -> Iterator[str]:
+    conversation = []
+    if system_prompt:
+        conversation.append({"role": "system", "content": system_prompt})
+    for user, assistant in chat_history:
+        conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
+    prompt = f"""### Instruction:
+                根据巨杉数据库SequoiaDB的相关问题进行回答。
+                
+                ### Input:
+                {message}
+                
+                ### Response:
             """
-    partial_message = ""
-    for token in client.text_generation(prompt=prompt, max_new_tokens=512, stream=True, best_of=1, temperature=0.1, 
-                                        top_p=0.99, do_sample=True, repetition_penalty=1.2):
-        if token.startswith("<s>"):
-            return partial_message
-        partial_message += token
-        yield partial_message
-
-gr.ChatInterface(
-    inference,
-    chatbot=gr.Chatbot(height=300, scale=7),
-    textbox=gr.Textbox(placeholder="你可以问我任何关于SequioaDB的问题!", container=False, scale=7),
-    description="这是SequioaDB旗下的AI智能大语言模型，训练超过上万条真实数据和7亿参数。",
-    title="ChatSDB",
-    examples=["SequoiaDB巨杉数据库是什么？", "SequoiaDB巨杉数据库支持哪些类型的数据库实例？"],
-    retry_btn="重试",
-    undo_btn="撤销",
-    clear_btn="清除",
-    submit_btn="提问",
-).queue().launch()
+    conversation.append({"role": "user", "content": prompt})
+    chat = tokenizer.apply_chat_template(conversation, tokenize=False)
+    inputs = tokenizer(chat, return_tensors="pt", truncation=True).input_ids.cuda()
+    if len(inputs) > MAX_INPUT_TOKEN_LENGTH:
+        inputs = inputs[-MAX_INPUT_TOKEN_LENGTH:]
+        gr.Warning("Trimmed input from conversation as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens.")
+
+    streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
+    generate_kwargs = dict(
+        inputs,
+        streamer=streamer,
+        max_new_tokens=max_new_tokens,
+        do_sample=True,
+        top_p=top_p,
+        top_k=top_k,
+        temperature=temperature,
+        num_beams=1,
+        repetition_penalty=repetition_penalty,
+    )
+    t = Thread(target=model.generate, kwargs=generate_kwargs)
+    t.start()
+
+    outputs = []
+    for text in streamer:
+        outputs.append(text)
+        yield "".join(outputs)
+
+
+chat_interface = gr.ChatInterface(
+    fn=generate,
+    additional_inputs=[
+        gr.Textbox(label="System prompt", lines=6),
+        gr.Slider(
+            label="Max new tokens",
+            minimum=1,
+            maximum=MAX_MAX_NEW_TOKENS,
+            step=1,
+            value=DEFAULT_MAX_NEW_TOKENS,
+        ),
+        gr.Slider(
+            label="Temperature",
+            minimum=0.1,
+            maximum=4.0,
+            step=0.1,
+            value=0.6,
+        ),
+        gr.Slider(
+            label="Top-p (nucleus sampling)",
+            minimum=0.05,
+            maximum=1.0,
+            step=0.05,
+            value=0.9,
+        ),
+        gr.Slider(
+            label="Top-k",
+            minimum=1,
+            maximum=1000,
+            step=1,
+            value=50,
+        ),
+        gr.Slider(
+            label="Repetition penalty",
+            minimum=1.0,
+            maximum=2.0,
+            step=0.05,
+            value=1.2,
+        ),
+    ],
+    stop_btn=None,
+    examples=[
+        ["如何安装巨杉数据库SequioaDB？"],
+        ["巨杉数据库SequioaDB有哪些优势？"],
+        ["巨杉数据库SequioaDB是什么？"],
+    ],
+)
+
+with gr.Blocks(css="style.css") as demo:
+    gr.Markdown(DESCRIPTION)
+    gr.DuplicateButton(value="Duplicate Space for private use", elem_id="duplicate-button")
+    chat_interface.render()
+    gr.Markdown(LICENSE)
+
+if __name__ == "__main__":
+    demo.queue(max_size=20).launch()