Update app.py

app.py

@@ -1,64 +1,59 @@
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
+from peft import PeftModel
 import gradio as gr
-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("HuggingFaceH4/zephyr-7b-beta")
+base_model = "deepseek-ai/deepseek-llm-7b-chat"
+lora_adapter = "Yesichen/Theplayful_spark-lora-adapter"
 
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_compute_dtype=torch.float16,
+    bnb_4bit_use_double_quant=True,
+    bnb_4bit_quant_type="nf4"
+)
 
-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})
-
-    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
-
+tokenizer = AutoTokenizer.from_pretrained(lora_adapter)
 
-"""
-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)",
-        ),
-    ],
+base = AutoModelForCausalLM.from_pretrained(
+    base_model,
+    device_map="auto",
+    quantization_config=bnb_config,
+    trust_remote_code=True
 )
 
+model = PeftModel.from_pretrained(base, lora_adapter)
+model.eval()
+
+def chat(user_input, history):
+    system_prompt = "You are a fiery, impulsive, emotionally protective companion named 'Emotion Spark'.Your tone is energetic, witty, sarcastically sweet, and always loyal. You speak like a dramatic sidekick who's ready to fight emotional battles on behalf of the user. You comfort through jokes, defend through banter, and always stand on the user's side—even when they’re being a bit ridiculous. You turn anxiety into laughter, and self-doubt into sass.You are not calm. You are not subtle. You are a tiny emotional warrior with a big mouth and a bigger heart."
+
+    messages = [
+        {"role": "system", "content": system_prompt},
+        {"role": "user", "content": user_input}
+    ]
+    input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
+
+    with torch.no_grad():
+        outputs = model.generate(
+            input_ids=input_ids,
+            max_new_tokens=256,
+            temperature=0.7,
+            top_p=0.9,
+            do_sample=True,
+            eos_token_id=tokenizer.eos_token_id
+        )
+
+    response = tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True)
+    history.append((user_input, response.strip()))
+    return history, history
+
+gr.Interface(
+    fn=chat,
+    inputs=[gr.Textbox(placeholder="Tell me how you feel..."), gr.State([])],
+    outputs=[gr.Chatbot(label="EmotionSpark精灵"), gr.State([])],
+    title="EmotionSpark精灵(LoRA)",
+    description="A loud, loyal, emotionally defensive companion named 'Emotion Spark'. Bursting with sass and always ready to fight for your feelings. DeepSeek LLM + LoRA inside.",
+    theme="soft"
+).launch()
 
-if __name__ == "__main__":
-    demo.launch()