Spaces:
Sleeping
Sleeping
Download app,py from abodealdrb/Qalam-University-API: direct link, hf CLI and curl.
- Browser
- Download file 2.47 kB
-
https://huggingface.co/spaces/abodealdrb/Qalam-University-API/resolve/main/app%2Cpy
- Command line
-
hf download hf://spaces/abodealdrb/Qalam-University-API/app,py
-
curl -L -o app,py https://huggingface.co/spaces/abodealdrb/Qalam-University-API/resolve/main/app%2Cpy
2.47 kB
| import gradio as gr | |
| from unsloth import FastLanguageModel | |
| from unsloth.chat_templates import get_chat_template | |
| from langchain_community.embeddings import HuggingFaceEmbeddings | |
| from langchain_community.vectorstores import FAISS | |
| from langchain.text_splitter import RecursiveCharacterTextSplitter | |
| import os | |
| # 1. إعدادات الموديل (استخدم الموديل المرفوع) | |
| # ملاحظة: في الـ Space المجاني (CPU) قد يكون التحميل بطيئاً جداً | |
| # الأفضل استخدام الموديل المدمج (Merged) الذي شرحته لك سابقاً إذا كنت ستستخدم CPU | |
| MODEL_NAME = "abodealdrb/Yemen-University-Assistant-Llama3" | |
| token = os.getenv("HF_TOKEN") # يجب وضع التوكن في إعدادات السبيس | |
| model, tokenizer = FastLanguageModel.from_pretrained( | |
| model_name = MODEL_NAME, | |
| max_seq_length = 2048, | |
| dtype = None, | |
| load_in_4bit = True, | |
| token = token, | |
| ) | |
| FastLanguageModel.for_inference(model) | |
| tokenizer = get_chat_template(tokenizer, chat_template = "llama-3.1") | |
| # 2. تجهيز RAG | |
| print("Building RAG...") | |
| with open("knowledge_base.txt", "r", encoding="utf-8") as f: | |
| text = f.read() | |
| text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50) | |
| docs = text_splitter.create_documents([text]) | |
| embeddings = HuggingFaceEmbeddings(model_name="intfloat/multilingual-e5-small") | |
| db = FAISS.from_documents(docs, embeddings) | |
| # 3. دالة الرد | |
| def predict(message): | |
| # بحث RAG | |
| context = "" | |
| if db: | |
| search_docs = db.similarity_search(message, k=3) | |
| context = "\n".join([d.page_content for d in search_docs]) | |
| # البرومبت | |
| system_prompt = f""" | |
| أنت مساعد جامعة القلم. | |
| استخدم السياق التالي للإجابة بدقة: | |
| {context} | |
| """ | |
| messages = [ | |
| {"role": "system", "content": system_prompt}, | |
| {"role": "user", "content": message} | |
| ] | |
| inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt") | |
| outputs = model.generate(input_ids=inputs, max_new_tokens=256, temperature=0.1) | |
| response = tokenizer.decode(outputs[0][inputs.shape[-1]:], skip_special_tokens=True) | |
| return response | |
| # 4. تشغيل التطبيق كـ API | |
| iface = gr.Interface(fn=predict, inputs="text", outputs="text") | |
| iface.launch() |