main.py

from flask import Flask, request, jsonify

device = "cuda" # the device to load the model onto

from ctransformers import AutoModelForCausalLM

llm = AutoModelForCausalLM.from_pretrained("TheBloke/Mistral-7B-v0.1-GGUF", model_file="mistral-7b-v0.1.Q4_K_M.gguf", model_type="mistral", gpu_layers=00)


@app.route('/recommend', methods=['POST'])
def recommendation():
    content = request.json
    user_degree = content.get('degree')
    user_stream = content.get('stream')
    user_semester = content.get('semester')
    prompt = """
    You need to act like as recommendataion engine for course recommendation based on below details.

    Degree: {user_degree}
    Stream: {user_stream}
    Current Semester: {user_semester}


    Based on above details recommend the courses that realtes to above details 
    Note: Output should bevalid json format in below format:
    {{"course1:ABC,course2:DEF,course3:XYZ,...}}
    
    """
    suffix="[/INST]"
    prefix="[INST] <<SYS>> You are a helpful assistant <</SYS>>"
    prompt = f"{prefix}{user.replace('{prompt}', prompt)}{suffix}"
    return jsonify({"ans":llm(prompt)})

if __name__ == '__main__':
    app.run(debug=True)