File size: 3,619 Bytes
81a49f7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
106
107
108
109

#Ajetaan tarvittavat kirjastot
import gradio as gr
import torch
import os
from huggingface_hub import login
from transformers import AutoTokenizer, AutoModelForCausalLM
from huggingface_hub import login

token = os.getenv("GreenerGlass")  # Get from Hugging Face secret
if token:
    login(token=token)

# Lataa malli tokenin kanssa
model_name = "google/gemma-2-2b-it" # Corrected model name based on previous successful load
tokenizer = AutoTokenizer.from_pretrained(model_name, token=token) # Pass token here
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    token=token, # Pass token here
    device_map="auto",
    torch_dtype=torch.float16
)
device = "cuda" if torch.cuda.is_available() else "cpu"

def generate_text(job_title, num_questions, temperature):
    # Muodosta prompt haastatttelukysymysten generointiin
    prompt = f"Generate {num_questions} professional interview questions for a {job_title} position. Provide clear, insightful questions that assess the candidate's skills and experience:"

    # Tokenize the input prompt
    inputs = tokenizer.encode(prompt, return_tensors="pt").to(model.device)

    # Generate text using the model
    outputs = model.generate(
        inputs,
        max_length=300,  # Riittävä pituus useammalle kysymykselle
        temperature=temperature,
        num_return_sequences=1,
        do_sample=True,
        top_p=0.9,
        top_k=50
    )
    # Decode the generated text
    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

    # Process the generated text to remove asterisks and extract questions
    processed_text = generated_text.replace('*', '') # Remove asterisks

    return processed_text

# Gradio custom theme
custom_theme = gr.themes.Base(
    primary_hue=gr.themes.Color(
        name="green",
        c50="#e8f5e9",
        c100="#c8e6c9",
        c200="#a5d6a7",
        c300="#81c784",
        c400="#66bb6a",
        c500="#4caf50",
        c600="#43a047",
        c700="#388e3c",
        c800="#2e7d32",
        c900="#1b5e20",
        c950="#0d3b0d", # Dark green for background
    ),
    neutral_hue=gr.themes.Color(
        name="gray",
        c50="#f9fafb",
        c100="#f3f4f6",
        c200="#e5e7eb",
        c300="#d1d5db",
        c400="#9ca3af",
        c500="#6b7280",
        c600="#4b5563",
        c700="#374151",
        c800="#1f2937",
        c900="#111827",
        c950="#030712",
    ),
).set(
    body_background_fill_dark="--primary-950", # Set body background to dark green
)

# Gradio-käyttöliittymä
with gr.Blocks(theme=custom_theme) as interface:
    gr.Markdown("🍀 **GreenerGlass question manager** 🍀")
    gr.Markdown("""Powered by Google's Gemma 2 model to generate professional interview questions.
    ⚠️ Note: Works better in English. ⚠️""")
    job_title_input = gr.Textbox(
        label="Job Title ",
        placeholder="Job title in English, e.g. Software Developer",
        lines=2
    )
    with gr.Accordion("More options"):
        num_questions_slider = gr.Slider(3, 8, value=5, step=1, label="Number of Questions")
        temperature_slider = gr.Slider(0.6, 1.2, value=0.8, step=0.1, label="Temperature (higher = more creative)")

    generate_button = gr.Button("Generate Questions", variant='primary') # Ensure variant is set to primary
    output_text = gr.Textbox(label="Interview Questions", lines=15)

    generate_button.click(
        fn=generate_text,
        inputs=[job_title_input, num_questions_slider, temperature_slider],
        outputs=output_text
    )


if __name__ == "__main__":
    interface.launch(share=True)