Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,234 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 2 |
+
from tokenization_yi import YiTokenizer
|
| 3 |
+
import torch
|
| 4 |
+
import os
|
| 5 |
+
import gradio as gr
|
| 6 |
+
import sentencepiece
|
| 7 |
+
|
| 8 |
+
model_id = "01-ai/Yi-34B-200K"
|
| 9 |
+
|
| 10 |
+
os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'max_split_size_mb:50'
|
| 11 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 12 |
+
offload_directory = './model_offload'
|
| 13 |
+
if not os.path.exists(offload_directory):
|
| 14 |
+
os.makedirs(offload_directory)
|
| 15 |
+
|
| 16 |
+
tokenizer = YiTokenizer(vocab_file="./tokenizer.model")
|
| 17 |
+
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_8bit=True, load_in_8bit_fp32_cpu_offload=True, offload_folder=offload_directory, trust_remote_code=True)
|
| 18 |
+
# model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True)
|
| 19 |
+
# model = model.to(device)
|
| 20 |
+
|
| 21 |
+
def run(message, chat_history, max_new_tokens=4056, temperature=3.5, top_p=0.9, top_k=800):
|
| 22 |
+
prompt = get_prompt(message, chat_history)
|
| 23 |
+
input_ids = tokenizer.encode(prompt, return_tensors='pt')
|
| 24 |
+
input_ids = input_ids.to(model.device)
|
| 25 |
+
response_ids = model.generate(
|
| 26 |
+
input_ids,
|
| 27 |
+
max_length=max_new_tokens + input_ids.shape[1],
|
| 28 |
+
temperature=temperature,
|
| 29 |
+
top_p=top_p,
|
| 30 |
+
top_k=top_k,
|
| 31 |
+
pad_token_id=tokenizer.eos_token_id,
|
| 32 |
+
do_sample=True
|
| 33 |
+
|
| 34 |
+
)
|
| 35 |
+
|
| 36 |
+
response = tokenizer.decode(response_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)
|
| 37 |
+
return response
|
| 38 |
+
|
| 39 |
+
def get_prompt(message, chat_history):
|
| 40 |
+
texts = []
|
| 41 |
+
|
| 42 |
+
do_strip = False
|
| 43 |
+
for user_input, response in chat_history:
|
| 44 |
+
user_input = user_input.strip() if do_strip else user_input
|
| 45 |
+
do_strip = True
|
| 46 |
+
texts.append(f" {response.strip()} {user_input} ")
|
| 47 |
+
message = message.strip() if do_strip else message
|
| 48 |
+
texts.append(f"{message}")
|
| 49 |
+
return ''.join(texts)
|
| 50 |
+
|
| 51 |
+
DESCRIPTION = """
|
| 52 |
+
# 👋🏻Welcome to 🙋🏻♂️Tonic's🧑🏻🚀YI-200K🚀"
|
| 53 |
+
You can use this Space to test out the current model [Tonic/YI](https://huggingface.co/01-ai/Yi-34B)
|
| 54 |
+
You can also use 🧑🏻🚀YI-200K🚀 by cloning this space. 🧬🔬🔍 Simply click here: <a style="display:inline-block" href="https://huggingface.co/spaces/Tonic1/YiTonic?duplicate=true"><img src="https://img.shields.io/badge/-Duplicate%20Space-blue?labelColor=white&style=flat&logo=data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/9hAAAAAXNSR0IArs4c6QAAAP5JREFUOE+lk7FqAkEURY+ltunEgFXS2sZGIbXfEPdLlnxJyDdYB62sbbUKpLbVNhyYFzbrrA74YJlh9r079973psed0cvUD4A+4HoCjsA85X0Dfn/RBLBgBDxnQPfAEJgBY+A9gALA4tcbamSzS4xq4FOQAJgCDwV2CPKV8tZAJcAjMMkUe1vX+U+SMhfAJEHasQIWmXNN3abzDwHUrgcRGmYcgKe0bxrblHEB4E/pndMazNpSZGcsZdBlYJcEL9Afo75molJyM2FxmPgmgPqlWNLGfwZGG6UiyEvLzHYDmoPkDDiNm9JR9uboiONcBXrpY1qmgs21x1QwyZcpvxt9NS09PlsPAAAAAElFTkSuQmCC&logoWidth=14" alt="Duplicate Space"></a></h3>
|
| 55 |
+
Join us : 🌟TeamTonic🌟 is always making cool demos! Join our active builder's🛠️community on 👻Discord: [Discord](https://discord.gg/nXx5wbX9) On 🤗Huggingface: [TeamTonic](https://huggingface.co/TeamTonic) & [MultiTransformer](https://huggingface.co/MultiTransformer) On 🌐Github: [Polytonic](https://github.com/tonic-ai) & contribute to 🌟 [PolyGPT](https://github.com/tonic-ai/polygpt-alpha)
|
| 56 |
+
"""
|
| 57 |
+
|
| 58 |
+
MAX_MAX_NEW_TOKENS = 4056
|
| 59 |
+
DEFAULT_MAX_NEW_TOKENS = 1256
|
| 60 |
+
MAX_INPUT_TOKEN_LENGTH = 120000
|
| 61 |
+
|
| 62 |
+
def clear_and_save_textbox(message): return '', message
|
| 63 |
+
|
| 64 |
+
def display_input(message, history=[]):
|
| 65 |
+
history.append((message, ''))
|
| 66 |
+
return history
|
| 67 |
+
|
| 68 |
+
def delete_prev_fn(history=[]):
|
| 69 |
+
try:
|
| 70 |
+
message, _ = history.pop()
|
| 71 |
+
except IndexError:
|
| 72 |
+
message = ''
|
| 73 |
+
return history, message or ''
|
| 74 |
+
|
| 75 |
+
def generate(message, history_with_input, max_new_tokens, temperature, top_p, top_k):
|
| 76 |
+
if int(max_new_tokens) > MAX_MAX_NEW_TOKENS:
|
| 77 |
+
raise ValueError
|
| 78 |
+
|
| 79 |
+
history = history_with_input[:-1]
|
| 80 |
+
response = run(message, history, max_new_tokens, temperature, top_p, top_k)
|
| 81 |
+
yield history + [(message, response)]
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
def process_example(message):
|
| 85 |
+
generator = generate(message, [], 1024, 2.5, 0.95, 900)
|
| 86 |
+
for x in generator:
|
| 87 |
+
pass
|
| 88 |
+
return '', x
|
| 89 |
+
|
| 90 |
+
def check_input_token_length(message, chat_history):
|
| 91 |
+
input_token_length = len(message) + len(chat_history)
|
| 92 |
+
if input_token_length > MAX_INPUT_TOKEN_LENGTH:
|
| 93 |
+
raise gr.Error(f"The accumulated input is too long ({input_token_length} > {MAX_INPUT_TOKEN_LENGTH}). Clear your chat history and try again.")
|
| 94 |
+
|
| 95 |
+
with gr.Blocks(theme='ParityError/Anime') as demo:
|
| 96 |
+
gr.Markdown(DESCRIPTION)
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
with gr.Group():
|
| 101 |
+
chatbot = gr.Chatbot(label='TonicYi-30B-200K')
|
| 102 |
+
with gr.Row():
|
| 103 |
+
textbox = gr.Textbox(
|
| 104 |
+
container=False,
|
| 105 |
+
show_label=False,
|
| 106 |
+
placeholder='As the dawn approached, they leant in and said',
|
| 107 |
+
scale=10
|
| 108 |
+
)
|
| 109 |
+
submit_button = gr.Button('Submit', variant='primary', scale=1, min_width=0)
|
| 110 |
+
|
| 111 |
+
with gr.Row():
|
| 112 |
+
retry_button = gr.Button('Retry', variant='secondary')
|
| 113 |
+
undo_button = gr.Button('Undo', variant='secondary')
|
| 114 |
+
clear_button = gr.Button('Clear', variant='secondary')
|
| 115 |
+
|
| 116 |
+
saved_input = gr.State()
|
| 117 |
+
|
| 118 |
+
with gr.Accordion(label='Advanced options', open=False):
|
| 119 |
+
# system_prompt = gr.Textbox(label='System prompt', value=DEFAULT_SYSTEM_PROMPT, lines=5, interactive=False)
|
| 120 |
+
max_new_tokens = gr.Slider(label='Max New Tokens', minimum=1, maximum=MAX_MAX_NEW_TOKENS, step=1, value=DEFAULT_MAX_NEW_TOKENS)
|
| 121 |
+
temperature = gr.Slider(label='Temperature', minimum=0.1, maximum=4.0, step=0.1, value=0.1)
|
| 122 |
+
top_p = gr.Slider(label='Top-P (nucleus sampling)', minimum=0.05, maximum=1.0, step=0.05, value=0.9)
|
| 123 |
+
top_k = gr.Slider(label='Top-K', minimum=1, maximum=1000, step=1, value=10)
|
| 124 |
+
|
| 125 |
+
textbox.submit(
|
| 126 |
+
fn=clear_and_save_textbox,
|
| 127 |
+
inputs=textbox,
|
| 128 |
+
outputs=[textbox, saved_input],
|
| 129 |
+
api_name=False,
|
| 130 |
+
queue=False,
|
| 131 |
+
).then(
|
| 132 |
+
fn=display_input,
|
| 133 |
+
inputs=[saved_input, chatbot],
|
| 134 |
+
outputs=chatbot,
|
| 135 |
+
api_name=False,
|
| 136 |
+
queue=False,
|
| 137 |
+
).then(
|
| 138 |
+
fn=check_input_token_length,
|
| 139 |
+
inputs=[saved_input, chatbot],
|
| 140 |
+
api_name=False,
|
| 141 |
+
queue=False,
|
| 142 |
+
).success(
|
| 143 |
+
fn=generate,
|
| 144 |
+
inputs=[
|
| 145 |
+
saved_input,
|
| 146 |
+
chatbot,
|
| 147 |
+
max_new_tokens,
|
| 148 |
+
temperature,
|
| 149 |
+
top_p,
|
| 150 |
+
top_k,
|
| 151 |
+
],
|
| 152 |
+
outputs=chatbot,
|
| 153 |
+
api_name="Generate",
|
| 154 |
+
)
|
| 155 |
+
|
| 156 |
+
button_event_preprocess = submit_button.click(
|
| 157 |
+
fn=clear_and_save_textbox,
|
| 158 |
+
inputs=textbox,
|
| 159 |
+
outputs=[textbox, saved_input],
|
| 160 |
+
api_name=False,
|
| 161 |
+
queue=False,
|
| 162 |
+
).then(
|
| 163 |
+
fn=display_input,
|
| 164 |
+
inputs=[saved_input, chatbot],
|
| 165 |
+
outputs=chatbot,
|
| 166 |
+
api_name=False,
|
| 167 |
+
queue=False,
|
| 168 |
+
).then(
|
| 169 |
+
fn=check_input_token_length,
|
| 170 |
+
inputs=[saved_input, chatbot],
|
| 171 |
+
api_name=False,
|
| 172 |
+
queue=False,
|
| 173 |
+
).success(
|
| 174 |
+
fn=generate,
|
| 175 |
+
inputs=[
|
| 176 |
+
saved_input,
|
| 177 |
+
chatbot,
|
| 178 |
+
max_new_tokens,
|
| 179 |
+
temperature,
|
| 180 |
+
top_p,
|
| 181 |
+
top_k,
|
| 182 |
+
],
|
| 183 |
+
outputs=chatbot,
|
| 184 |
+
api_name="Cgenerate",
|
| 185 |
+
)
|
| 186 |
+
|
| 187 |
+
retry_button.click(
|
| 188 |
+
fn=delete_prev_fn,
|
| 189 |
+
inputs=chatbot,
|
| 190 |
+
outputs=[chatbot, saved_input],
|
| 191 |
+
api_name=False,
|
| 192 |
+
queue=False,
|
| 193 |
+
).then(
|
| 194 |
+
fn=display_input,
|
| 195 |
+
inputs=[saved_input, chatbot],
|
| 196 |
+
outputs=chatbot,
|
| 197 |
+
api_name=False,
|
| 198 |
+
queue=False,
|
| 199 |
+
).then(
|
| 200 |
+
fn=generate,
|
| 201 |
+
inputs=[
|
| 202 |
+
saved_input,
|
| 203 |
+
chatbot,
|
| 204 |
+
max_new_tokens,
|
| 205 |
+
temperature,
|
| 206 |
+
top_p,
|
| 207 |
+
top_k,
|
| 208 |
+
],
|
| 209 |
+
outputs=chatbot,
|
| 210 |
+
api_name=False,
|
| 211 |
+
)
|
| 212 |
+
|
| 213 |
+
undo_button.click(
|
| 214 |
+
fn=delete_prev_fn,
|
| 215 |
+
inputs=chatbot,
|
| 216 |
+
outputs=[chatbot, saved_input],
|
| 217 |
+
api_name=False,
|
| 218 |
+
queue=False,
|
| 219 |
+
).then(
|
| 220 |
+
fn=lambda x: x,
|
| 221 |
+
inputs=[saved_input],
|
| 222 |
+
outputs=textbox,
|
| 223 |
+
api_name=False,
|
| 224 |
+
queue=False,
|
| 225 |
+
)
|
| 226 |
+
|
| 227 |
+
clear_button.click(
|
| 228 |
+
fn=lambda: ([], ''),
|
| 229 |
+
outputs=[chatbot, saved_input],
|
| 230 |
+
queue=False,
|
| 231 |
+
api_name=False,
|
| 232 |
+
)
|
| 233 |
+
|
| 234 |
+
demo.queue().launch(show_api=True)
|