File size: 2,749 Bytes
60a9633
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from conversation import get_default_conv_template
import base64
import streamlit as st

def generate(style, topic, words=200, sender='Sender_Name', recipient='Recipient_Name'):
    tokenizer = AutoTokenizer.from_pretrained("GeneZC/MiniChat-3B", use_fast=False)
    device = 'cuda' if torch.cuda.is_available() else 'cpu'
    def get_model(device):
        if device == 'cuda':
            return AutoModelForCausalLM.from_pretrained("GeneZC/MiniChat-3B", use_cache=True, device_map="auto", torch_dtype=torch.float16).eval()
        else:
            return AutoModelForCausalLM.from_pretrained("GeneZC/MiniChat-3B", use_cache=True, device_map="cpu", torch_dtype=torch.float32).eval()

    model = get_model(device)
    conv = get_default_conv_template("minichat")

    question = f"""
        Generate an email with the following specifications and make sure to highlight the subject, according to the topic of the mail, in the very first line of the response:
      - Style: {style}
      - Word Limit: {words}
      - Topic: {topic}
      - Sender Name: {sender}
      - Recipient Name: {recipient}
    """

    conv.append_message(conv.roles[0], question)
    conv.append_message(conv.roles[1], None)
    prompt = conv.get_prompt()
    input_ids = tokenizer([prompt]).input_ids
    output_ids = model.generate(
        torch.as_tensor(input_ids).to(device),
        do_sample=True,
        temperature=0.7,
        max_new_tokens=2000,
    )
    output_ids = output_ids[0][len(input_ids[0]):]
    output = tokenizer.decode(output_ids, skip_special_tokens=True).strip()
    return output

@st.cache_data
def get_image(file):
  with open(file, "rb") as f:
      data = f.read()
  return base64.b64encode(data).decode()

with open("./static/style.css") as style:
    st.markdown(f"<style>{style.read()}</style>", unsafe_allow_html=True)

img = get_image("./static/bg.png")
bg_image = f"""
<style>
[data-testid="stAppViewContainer"]{{
    background-image: url("data:image/png;base64,{img}");
    background-size: cover;
}}
</style>
"""
st.markdown(bg_image, unsafe_allow_html=True)
st.title("MailCraft")
st.subheader("Unlock Seamless Email Excellence 📧 for Effortless Communication")

style = st.text_input("Enter Email Type", placeholder="Ex. Professional/Personal/Job Application")
words = st.number_input("Enter Word Limit", min_value=100, max_value=500, value=None, placeholder="From 100 to 500")
topic = st.text_input("Enter Email Topic")
sender = st.text_input("Enter Sender Name")
recipient = st.text_input("Enter Recipient Name")

if st.button("Generate Email"):
    generated_email = generate(style, topic, words, sender, recipient)
    st.write(generated_email)