File size: 5,582 Bytes
c544e81
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
49151d9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c544e81
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
from huggingface_hub import InferenceClient
import gradio as gr

client = InferenceClient(
    "mistralai/Mistral-7B-Instruct-v0.1"
)


def format_prompt(message, history):
  prompt = "<s>"
  for user_prompt, bot_response in history:
    prompt += f"[INST] {user_prompt} [/INST]"
    prompt += f" {bot_response}</s> "
  prompt += f"[INST] {message} [/INST]"
  return prompt

def generate(
    prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
):
    temperature = float(temperature)
    if temperature < 1e-2:
        temperature = 1e-2
    top_p = float(top_p)

    generate_kwargs = dict(
        temperature=temperature,
        max_new_tokens=max_new_tokens,
        top_p=top_p,
        repetition_penalty=repetition_penalty,
        do_sample=True,
        seed=42,
    )

    formatted_prompt = format_prompt(prompt, history)

    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
    output = ""

    for response in stream:
        output += response.token.text
        yield output
    return output


additional_inputs=[
    gr.Slider(
        label="Temperature",
        value=0.9,
        minimum=0.0,
        maximum=1.0,
        step=0.05,
        interactive=True,
        info="Higher values produce more diverse outputs",
    ),
    gr.Slider(
        label="Max new tokens",
        value=256,
        minimum=0,
        maximum=1048,
        step=64,
        interactive=True,
        info="The maximum numbers of new tokens",
    ),
    gr.Slider(
        label="Top-p (nucleus sampling)",
        value=0.90,
        minimum=0.0,
        maximum=1,
        step=0.05,
        interactive=True,
        info="Higher values sample more low-probability tokens",
    ),
    gr.Slider(
        label="Repetition penalty",
        value=1.2,
        minimum=1.0,
        maximum=2.0,
        step=0.05,
        interactive=True,
        info="Penalize repeated tokens",
    )
]

css = """
  #mkd {
    height: 200px; 
    overflow: auto; 
    border: 1px solid #ccc; 
  }
"""

with gr.Blocks(css=css) as demo:
  
    gr.ChatInterface(
        generate,
        additional_inputs=additional_inputs,
        examples = [
            ["๐Ÿฐ Welcome to the Kingdom of Elandria! You are Jim and Tim, two bumbling bros with a knack for mischief. ๐Ÿคด๐Ÿคด [Action: Introduce yourselves, Equipment: Scepters of Foolishness]"],
            ["๐ŸŒฒ You find yourselves in a forest filled with magical creatures and oddly specific 'Do Not Disturb' signs. ๐Ÿฆ„ [Action: Proceed cautiously, Equipment: Map of Social Etiquette]"],
            ["๐Ÿป You stumble upon a dwarf tavern. They offer you 'Beard Beer.' Do you drink it? ๐Ÿบ [Action: Chug or Pass, Equipment: Empty Mugs]"],
            ["๐Ÿ‰ A vegan dragon appears and chastises you for your leather boots. What do you do? ๐Ÿ—ก๏ธ๐Ÿƒ [Action: Apologize and offer kale, Equipment: Non-leather sandals]"],
            ["๐Ÿ’Ž You find a treasure chest labeled 'Not a Mimic.' Seems legit. Do you open it? ๐Ÿ—๏ธ [Action: Open or No way, Equipment: Mimic Repellent]"],
            ["๐Ÿฆ‡ You're swarmed by bats in a cave. One bat offers you 'organic guano.' How do you react? ๐Ÿ•ฏ๏ธ [Action: Politely decline, Equipment: Nose Plugs]"],
            ["๐Ÿ”ฎ A mysterious sorcerer offers you a 'Love Potion No. 9ยฝ.' Do you take a sip? ๐Ÿถ [Action: Sip or Skip, Equipment: Breath Mints]"],
            ["โš”๏ธ Bandits demand gold, but they accept credit cards. What's your move? ๐Ÿ’ฐ [Action: Pay or Pray, Equipment: Wallets]"],
            ["๐Ÿšช A door with three locks and a sign saying 'Beware of the Dog.' Do you search for the keys or try to pet the dog? ๐Ÿ—๏ธ๐Ÿ’ช [Action: Unlock or Pet, Equipment: Dog Treats]"],
            ["๐ŸŒŠ A river blocks your path. A mermaid offers to carry you across for a 'small' fee. ๐ŸŠโ€โ™€๏ธ๐ŸŒ‰ [Action: Accept or Decline, Equipment: Bargaining Skills]"],
            ["๐Ÿฆ You encounter a pride of lions playing poker. Do you join the game or fold? ๐Ÿคซ๐Ÿ”„ [Action: Play or Fold, Equipment: Poker Face]"],
            ["๐ŸŽ A tree filled with golden apples and a sign saying, 'Seriously, don't eat these!' What do you do? ๐Ÿค” [Action: Eat or Retreat, Equipment: Stomach Pump]"],
            ["๐ŸŒ• The moon turns red, wolves start howling, and your horoscope says 'Stay in bed.' Do you camp or go? ๐Ÿ•๏ธ๐Ÿšถ [Action: Camp or Scamp, Equipment: Astrology App]"],
            ["๐Ÿ’€ The final boss is an undead warrior selling life insurance. Do you combat or sign up? โš”๏ธ๐Ÿค [Action: Fight or Finance, Equipment: Policy Guide]"]
        ]
    )
    gr.HTML("""<h2>๐Ÿค– Mistral Chat - Gradio ๐Ÿค–</h2>
        In this demo, you can chat with <a href='https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1'>Mistral-7B-Instruct</a> model. ๐Ÿ’ฌ
        Learn more about the model <a href='https://huggingface.co/docs/transformers/main/model_doc/mistral'>here</a>. ๐Ÿ“š
        <h2>๐Ÿ›  Model Features ๐Ÿ› </h2>
        <ul>
          <li>๐ŸชŸ Sliding Window Attention with 128K tokens span</li>
          <li>๐Ÿš€ GQA for faster inference</li>
          <li>๐Ÿ“ Byte-fallback BPE tokenizer</li>
        </ul>
        <h3>๐Ÿ“œ License ๐Ÿ“œ  Released under Apache 2.0 License</h3>
        <h3>๐Ÿ“ฆ Usage ๐Ÿ“ฆ</h3>
        <ul>
          <li>๐Ÿ“š Available on Huggingface Hub</li>
          <li>๐Ÿ Python code snippets for easy setup</li>
          <li>๐Ÿ“ˆ Expected speedups with Flash Attention 2</li>
        </ul>
    """)
demo.queue().launch(debug=True)