File size: 4,437 Bytes
bd0305e
 
015885c
bd0305e
 
 
d14c800
015885c
 
1019a35
d14c800
 
 
1019a35
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c58f313
 
14e5da3
c58f313
 
 
 
 
015885c
14e5da3
 
d14c800
 
 
c58f313
14e5da3
015885c
c58f313
 
015885c
 
 
 
 
 
 
 
 
 
 
 
 
 
c58f313
14e5da3
 
015885c
c58f313
14e5da3
 
 
c58f313
 
1019a35
 
 
c58f313
1019a35
6badf1c
d14c800
 
c58f313
 
 
 
 
 
 
 
d14c800
 
1019a35
c58f313
 
 
 
 
 
 
14e5da3
c58f313
 
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
import gradio as gr
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline, StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer
import time
import numpy as np
from torch.nn import functional as F
import os
from threading import Thread

print(f"Starting to load the model to memory")
m = AutoModelForCausalLM.from_pretrained(
    "stabilityai/stablelm-tuned-alpha-7b", torch_dtype=torch.float16).cuda()
tok = AutoTokenizer.from_pretrained("stabilityai/stablelm-tuned-alpha-7b")
generator = pipeline('text-generation', model=m, tokenizer=tok, device=0)
print(f"Sucessfully loaded the model to the memory")

start_message = """<|SYSTEM|># StableAssistant
- StableAssistant is A helpful and harmless Open Source AI Language Model developed by Stability and CarperAI.
- StableAssistant is excited to be able to help the user, but will refuse to do anything that could be considered harmful to the user.
- StableAssistant is more than just an information source, StableAssistant is also able to write poetry, short stories, and make jokes.
- StableAssistant will refuse to participate in anything that could harm a human."""


class StopOnTokens(StoppingCriteria):
    def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
        stop_ids = [50278, 50279, 50277, 1, 0]
        for stop_id in stop_ids:
            if input_ids[0][-1] == stop_id:
                return True
        return False


def user(message, history):
    # Append the user's message to the conversation history
    return "", history + [[message, ""]]


def chat(curr_system_message, history):
    # Initialize a StopOnTokens object
    stop = StopOnTokens()

    # Construct the input message string for the model by concatenating the current system message and conversation history
    messages = curr_system_message + \
        "".join(["".join(["<|USER|>"+item[0], "<|ASSISTANT|>"+item[1]])
                for item in history])

    # Tokenize the messages string
    model_inputs = tok([messages], return_tensors="pt").to("cuda")
    streamer = TextIteratorStreamer(
        tok, timeout=10., skip_prompt=True, skip_special_tokens=True)
    generate_kwargs = dict(
        model_inputs,
        streamer=streamer,
        max_new_tokens=1024,
        do_sample=True,
        top_p=0.95,
        top_k=1000,
        temperature=1.0,
        num_beams=1,
        stopping_criteria=StoppingCriteriaList([stop])
    )
    t = Thread(target=m.generate, kwargs=generate_kwargs)
    t.start()

    # print(history)
    # Initialize an empty string to store the generated text
    partial_text = ""
    for new_text in streamer:
        # print(new_text)
        partial_text += new_text
        history[-1][1] = partial_text
        # Yield an empty string to cleanup the message textbox and the updated conversation history
        yield history
    return partial_text


with gr.Blocks() as demo:
    # history = gr.State([])
    gr.Markdown("## StableLM-Tuned-Alpha-7b Chat")
    gr.HTML('''<center><a href="https://huggingface.co/spaces/stabilityai/stablelm-tuned-alpha-chat?duplicate=true"><img src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>Duplicate the Space to skip the queue and run in a private space</center>''')
    chatbot = gr.Chatbot().style(height=500)
    with gr.Row():
        with gr.Column():
            msg = gr.Textbox(label="Chat Message Box", placeholder="Chat Message Box",
                             show_label=False).style(container=False)
        with gr.Column():
            with gr.Row():
                submit = gr.Button("Submit")
                stop = gr.Button("Stop")
                clear = gr.Button("Clear")
    system_msg = gr.Textbox(
        start_message, label="System Message", interactive=False, visible=False)

    submit_event = msg.submit(fn=user, inputs=[msg, chatbot], outputs=[msg, chatbot], queue=False).then(
        fn=chat, inputs=[system_msg, chatbot], outputs=[chatbot], queue=True)
    submit_click_event = submit.click(fn=user, inputs=[msg, chatbot], outputs=[msg, chatbot], queue=False).then(
        fn=chat, inputs=[system_msg, chatbot], outputs=[chatbot], queue=True)
    stop.click(fn=None, inputs=None, outputs=None, cancels=[
               submit_event, submit_click_event], queue=False)
    clear.click(lambda: None, None, [chatbot], queue=False)

demo.queue(max_size=32, concurrency_count=2)
demo.launch()