File size: 2,337 Bytes
486ca0f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
187e5b8
486ca0f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import gradio as gr
import torch
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    BitsAndBytesConfig,
    pipeline
)

model_name = "RaviNaik/Phi2-Osst"

device = "cuda" if torch.cuda.is_available() else "cpu"
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    device_map=device
)
model.config.use_cache = False
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True, device_map=device)
tokenizer.pad_token = tokenizer.eos_token
chat_template = """<|im_start|>system
You are a helpful assistant who always respond to user queries<|im_end|>
<im_start>user
{prompt}<|im_end|>
<|im_start|>assistant
"""

def generate(prompt, max_length, temperature, num_samples):
    prompt = prompt.strip()
    pipe = pipeline(task="text-generation", model=model, tokenizer=tokenizer, max_length=max_length, temperature=temperature, num_return_sequences=num_samples)
    # result = pipe(chat_template.format(prompt=prompt))
    result = pipe(prompt)
    return {output: result}


with gr.Blocks() as app:
    gr.Markdown("## ERA Session27 - Phi2 Model Finetuning with QLoRA on OpenAssistant Conversations Dataset (OASST1)")

    with gr.Row():
        with gr.Column():
            prompt_box = gr.Textbox(label="Initial Prompt", interactive=True)
            max_length = gr.Slider(
                minimum=50,
                maximum=500,
                value=200,
                step=10,
                label="Select Number of Tokens to be Generated",
                interactive=True,
            )
            temperature = gr.Slider(
                minimum=0.1,
                maximum=1,
                value=0.7,
                step=0.1,
                label="Select Temperature",
                interactive=True,
            )
            num_samples = gr.Dropdown(
                choices=[1, 2, 5, 10],
                value=1,
                interactive=True,
                label="Select No. of outputs to be generated",
            )
            submit_btn = gr.Button(value="Generate")

        with gr.Column():
            output = gr.JSON(label="Generated Text")

        submit_btn.click(
            generate,
            inputs=[prompt_box, max_length, temperature, num_samples],
            outputs=[output],
        )

app.launch()