Spaces:
Sleeping
Sleeping
File size: 3,087 Bytes
1fcd25d |
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 |
from flask import request,jsonify,Flask
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
app = Flask(__name__)
path_model ="./"
tokenizer = AutoTokenizer.from_pretrained(path_model)
model = AutoModelForCausalLM.from_pretrained(path_model, torch_dtype=torch.float32)
def generate_qa_summary(topic, num_questions, temperature=0.4):
device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)
questions_and_answers = []
while len(questions_and_answers) < num_questions:
prompt = f"{topic} Question:\n"
input_ids = tokenizer.encode(prompt, return_tensors="pt").to(device)
output = model.generate(
max_new_tokens=1000,
#length_penalty=1.0,
early_stopping=True,
input_ids=input_ids,
num_return_sequences=1,
temperature=temperature,
no_repeat_ngram_size=4,
pad_token_id=tokenizer.eos_token_id,
top_p=0.80,
do_sample=True,
)
response = tokenizer.decode(output[0], skip_special_tokens=True)
# Check if the response is valid (contains both 'Question:' and 'Answer:')
if "Question:" in response and "Answer:" in response:
questions_and_answers.append(response)
return questions_and_answers
@app.route("/api/Gen", methods=["POST"])
def generate_questions1():
data = request.get_json()
course_name = data["courseName"]
num_questions = int(data["numQuestions"])
qa_summary = generate_qa_summary(course_name, num_questions)
if not qa_summary:
return jsonify({"error": "Failed to generate any questions"})
first_item = qa_summary[0]
topic, _ = (
first_item.split("Question:", maxsplit=1)
if "Question:" in first_item
else (first_item, "")
)
topic = topic.strip()
formatted_summaries = [f"<strong>{topic}:</strong><br><br>"] # Start with the topic
for index, item in enumerate(qa_summary):
if "Question:" not in item:
item = f"Question: {item}" # Prepend "Question:" if missing
else:
parts = item.split("Question:")
item = "Question:" + " ".join(
parts[1:]
) # Reassemble without extra "Question:"
_, question_answer = item.split("Question:", maxsplit=1)
question, answer = (
question_answer.split("Answer:", maxsplit=1)
if "Answer:" in question_answer
else (question_answer, "No answer provided.")
)
formatted_question_answer = f"<div class='question-container'><div>-Question: {question.strip()}<br<button onclick='toggleAnswer({index})'><i id='icon{index}' class='fas fa-eye fa'></i></button></div><div id='answer{index}' style='display:none;'>-Answer: {answer.strip()}</div></div><br></div>"
formatted_summaries.append(formatted_question_answer)
return jsonify(formatted_summaries)
if __name__ == "__main__":
app.run(debug=True)
|