Spaces:
Sleeping
Sleeping
File size: 3,090 Bytes
e93a838 624da4f c730b5d 624da4f c730b5d 624da4f c730b5d 0c4938b 624da4f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 |
import os
from huggingface_hub import login
login(token=os.environ["HF_TOKEN"])
import warnings
warnings.filterwarnings('ignore')
import transformers
import gradio as gr
from transformers import pipeline
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_id = "meta-llama/Llama-2-7b-chat-hf"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
device_map="auto",
torch_dtype="auto"
)
# تحميل LLM آخر للتلخيص
summarizer =pipeline("summarization", model="facebook/bart-large-cnn")
def generate_response(user_input: str, context_summary: str = "") -> str:
prompt = f"""<s>[INST] <<SYS>>
You are a helpful assistant.
<</SYS>>
Previous summary of conversation:
{context_summary}
Current user input:
{user_input}
[/INST]
"""
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
output = model.generate(
**inputs,
max_new_tokens=256,
pad_token_id=tokenizer.eos_token_id,
temperature=0.7, # ← يتحكم في عشوائية الردود
top_p=0.9, # ← خوارزمية nucleus sampling
do_sample=True # ← فعّل العشوائية
)
response = tokenizer.decode(output[0], skip_special_tokens=True)
return response[len(prompt):].strip()
# تعريف chat_history كمخزن داخلي
chat_history = []
def reset_chat():
global chat_history
chat_history.clear()
return []
def gradio_chat(user_input):
global chat_history # نُشير إلى المتغيّر الخارجي
# استخراج الملخص السابق إن وجد
context_summary = chat_history[-1]["summary"] if chat_history else ""
# توليد الرد من الموديل
response = generate_response(user_input, context_summary)
# تلخيص الرد
summary = summarizer(
response,
max_length=100,
min_length=30,
do_sample=False
)[0]["summary_text"]
# حفظ في الذاكرة
chat_history.append({
"user_input": user_input,
"response": response,
"summary": summary
})
# تجهيز الشكل لعرض المحادثة
chat_display = [(entry["user_input"], entry["response"]) for entry in chat_history]
return chat_display
# واجهة Gradio
with gr.Blocks() as demo:
gr.Markdown("## 🤖 AI Chatbot with Memory\nKeep chatting and watch it remember!")
chatbot = gr.Chatbot()
msg = gr.Textbox(placeholder="Type your message and press Enter...")
clear_btn = gr.Button("🔄 Reset Chat")
# عند الإرسال، تُنفذ الدالة وتُحدث الـchatbot
msg.submit(fn=gradio_chat, inputs=msg, outputs=chatbot, scroll_to_output=True)
# بعد الإرسال، يتم مسح الـTextbox تلقائيًا
msg.submit(lambda: "", None, msg)
# زر إعادة تعيين المحادثة
clear_btn.click(fn=reset_chat, inputs=None, outputs=None)
# تشغيل الواجهة
demo.launch() |