Spaces:
Running
on
Zero
Running
on
Zero
Feature(MInference): add local mode
Browse files- app.py +90 -97
- requirements.txt +1 -2
app.py
CHANGED
@@ -1,53 +1,47 @@
|
|
1 |
import subprocess
|
2 |
import os
|
3 |
-
|
4 |
-
|
5 |
-
# os.environ["CPATH"] = "$CPATH:/usr/local/cuda/include"
|
6 |
-
# os.environ["CUDA_HOME"] = "/usr/local/cuda"
|
7 |
-
# os.environ["PATH"] = "$PATH:$CUDA_HOME/bin"
|
8 |
-
# os.environ["LIBRARY_PATH"] = "$LIBRARY_PATH:/usr/local/cuda/lib64"
|
9 |
-
|
10 |
-
subprocess.run(
|
11 |
-
"/home/user/.pyenv/shims/pip install flash-attn --no-build-isolation",
|
12 |
-
env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
|
13 |
-
shell=True,
|
14 |
-
)
|
15 |
-
# subprocess.run(
|
16 |
-
# "/home/user/.pyenv/shims/pip install pycuda==2023.1",
|
17 |
-
# env={"CPATH": "$CPATH:/usr/local/cuda/include", "LIBRARY_PATH": "$LIBRARY_PATH:/usr/local/cuda/lib64"},
|
18 |
-
# shell=True,
|
19 |
-
# )
|
20 |
|
21 |
import gradio as gr
|
22 |
import os
|
23 |
import spaces
|
24 |
-
from transformers import AutoModelForCausalLM
|
25 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
|
26 |
from threading import Thread
|
|
|
27 |
|
28 |
# Set an environment variable
|
29 |
HF_TOKEN = os.environ.get("HF_TOKEN", None)
|
30 |
|
31 |
|
32 |
-
DESCRIPTION =
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
<
|
37 |
-
<
|
38 |
-
</
|
39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
|
41 |
LICENSE = """
|
42 |
-
<
|
43 |
-
|
44 |
-
|
45 |
"""
|
46 |
|
47 |
PLACEHOLDER = """
|
48 |
<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
|
49 |
-
<
|
50 |
-
<h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">Meta llama3</h1>
|
51 |
<p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Ask me anything...</p>
|
52 |
</div>
|
53 |
"""
|
@@ -58,34 +52,28 @@ h1 {
|
|
58 |
text-align: center;
|
59 |
display: block;
|
60 |
}
|
61 |
-
#duplicate-button {
|
62 |
-
margin: auto;
|
63 |
-
color: white;
|
64 |
-
background: #1565c0;
|
65 |
-
border-radius: 100vh;
|
66 |
-
}
|
67 |
"""
|
68 |
|
69 |
# Load the tokenizer and model
|
70 |
-
model_name = "gradientai/Llama-3-8B-Instruct-
|
71 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
72 |
-
model = AutoModelForCausalLM.from_pretrained(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
|
74 |
-
|
75 |
-
minference_patch = MInference("minference", model_name)
|
76 |
-
model = minference_patch(model)
|
77 |
|
78 |
-
terminators = [
|
79 |
-
tokenizer.eos_token_id,
|
80 |
-
tokenizer.convert_tokens_to_ids("<|eot_id|>")
|
81 |
-
]
|
82 |
|
83 |
@spaces.GPU(duration=120)
|
84 |
-
def chat_llama3_8b(
|
85 |
-
|
86 |
-
|
87 |
-
max_new_tokens: int
|
88 |
-
) -> str:
|
89 |
"""
|
90 |
Generate a streaming response using the llama3-8b model.
|
91 |
Args:
|
@@ -99,81 +87,86 @@ def chat_llama3_8b(message: str,
|
|
99 |
# global model
|
100 |
conversation = []
|
101 |
for user, assistant in history:
|
102 |
-
conversation.extend(
|
|
|
|
|
|
|
|
|
|
|
103 |
conversation.append({"role": "user", "content": message})
|
104 |
|
105 |
-
input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
# from minference import MInference
|
113 |
-
# minference_patch = MInference("minference", model_name)
|
114 |
-
# model = minference_patch(model)
|
115 |
-
|
116 |
-
streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
|
117 |
|
118 |
generate_kwargs = dict(
|
119 |
-
input_ids=
|
120 |
streamer=streamer,
|
121 |
max_new_tokens=max_new_tokens,
|
122 |
do_sample=True,
|
123 |
temperature=temperature,
|
124 |
eos_token_id=terminators,
|
125 |
)
|
126 |
-
# This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash.
|
127 |
if temperature == 0:
|
128 |
-
generate_kwargs[
|
129 |
-
|
130 |
t = Thread(target=model.generate, kwargs=generate_kwargs)
|
131 |
t.start()
|
132 |
|
133 |
outputs = []
|
134 |
for text in streamer:
|
135 |
outputs.append(text)
|
136 |
-
#print(outputs)
|
137 |
yield "".join(outputs)
|
138 |
-
|
139 |
|
140 |
# Gradio block
|
141 |
-
chatbot=gr.Chatbot(height=450, placeholder=PLACEHOLDER, label=
|
142 |
|
143 |
with gr.Blocks(fill_height=True, css=css) as demo:
|
144 |
-
|
145 |
gr.Markdown(DESCRIPTION)
|
146 |
-
gr.DuplicateButton(value="Duplicate Space for private use", elem_id="duplicate-button")
|
147 |
gr.ChatInterface(
|
148 |
fn=chat_llama3_8b,
|
149 |
chatbot=chatbot,
|
150 |
fill_height=True,
|
151 |
-
additional_inputs_accordion=gr.Accordion(
|
|
|
|
|
152 |
additional_inputs=[
|
153 |
-
gr.Slider(
|
154 |
-
|
155 |
-
|
156 |
-
|
157 |
-
|
158 |
-
|
159 |
-
|
160 |
-
|
161 |
-
|
162 |
-
|
163 |
-
|
164 |
-
|
165 |
-
|
|
|
|
|
|
|
|
|
166 |
examples=[
|
167 |
-
[
|
168 |
-
[
|
169 |
-
[
|
170 |
-
[
|
171 |
-
[
|
172 |
-
|
173 |
cache_examples=False,
|
174 |
-
|
175 |
-
|
176 |
gr.Markdown(LICENSE)
|
177 |
-
|
178 |
if __name__ == "__main__":
|
179 |
-
demo.launch()
|
|
|
1 |
import subprocess
|
2 |
import os
|
3 |
+
import torch
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
|
5 |
import gradio as gr
|
6 |
import os
|
7 |
import spaces
|
|
|
8 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
|
9 |
from threading import Thread
|
10 |
+
from transformers.utils.import_utils import _is_package_available
|
11 |
|
12 |
# Set an environment variable
|
13 |
HF_TOKEN = os.environ.get("HF_TOKEN", None)
|
14 |
|
15 |
|
16 |
+
DESCRIPTION = """
|
17 |
+
# MInference 1.0: Accelerating Pre-filling for Long-Context LLMs via Dynamic Sparse Attention (Under Review) [[paper](https://arxiv.org/abs/2406.05736)]
|
18 |
+
_Huiqiang Jiang†, Yucheng Li†, Chengruidong Zhang†, Qianhui Wu, Xufang Luo, Surin Ahn, Zhenhua Han, Amir H. Abdi, Dongsheng Li, Chin-Yew Lin, Yuqing Yang and Lili Qiu_
|
19 |
+
|
20 |
+
<h2 style="text-align: center;"><a href="https://github.com/microsoft/MInference" target="blank"> [Code]</a>
|
21 |
+
<a href="https://hqjiang.com/minference.html" target="blank"> [Project Page]</a>
|
22 |
+
<a href="https://arxiv.org/abs/2406.05736" target="blank"> [Paper]</a></h2>
|
23 |
+
|
24 |
+
<font color="brown"><b>This is only a deployment demo. Due to limited GPU resources, we do not provide an online demo. You will need to follow the code below to try MInference locally.</b></font>
|
25 |
+
|
26 |
+
```bash
|
27 |
+
git clone https://huggingface.co/spaces/microsoft/MInference
|
28 |
+
cd MInference
|
29 |
+
pip install -r requirments.txt
|
30 |
+
pip install flash_attn pycuda==2023.1
|
31 |
+
python app.py
|
32 |
+
```
|
33 |
+
<br/>
|
34 |
+
"""
|
35 |
|
36 |
LICENSE = """
|
37 |
+
<div style="text-align: center;">
|
38 |
+
<p>© 2024 Microsoft</p>
|
39 |
+
</div>
|
40 |
"""
|
41 |
|
42 |
PLACEHOLDER = """
|
43 |
<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
|
44 |
+
<h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">LLaMA-3-8B-Gradient-1M w/ MInference</h1>
|
|
|
45 |
<p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Ask me anything...</p>
|
46 |
</div>
|
47 |
"""
|
|
|
52 |
text-align: center;
|
53 |
display: block;
|
54 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
55 |
"""
|
56 |
|
57 |
# Load the tokenizer and model
|
58 |
+
model_name = "gradientai/Llama-3-8B-Instruct-Gradient-1048k"
|
59 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
60 |
+
model = AutoModelForCausalLM.from_pretrained(
|
61 |
+
model_name, torch_dtype="auto", device_map="auto"
|
62 |
+
) # to("cuda:0")
|
63 |
+
|
64 |
+
if torch.cuda.is_available() and _is_package_available("pycuda"):
|
65 |
+
from minference import MInference
|
66 |
+
|
67 |
+
minference_patch = MInference("minference", model_name)
|
68 |
+
model = minference_patch(model)
|
69 |
|
70 |
+
terminators = [tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<|eot_id|>")]
|
|
|
|
|
71 |
|
|
|
|
|
|
|
|
|
72 |
|
73 |
@spaces.GPU(duration=120)
|
74 |
+
def chat_llama3_8b(
|
75 |
+
message: str, history: list, temperature: float, max_new_tokens: int
|
76 |
+
) -> str:
|
|
|
|
|
77 |
"""
|
78 |
Generate a streaming response using the llama3-8b model.
|
79 |
Args:
|
|
|
87 |
# global model
|
88 |
conversation = []
|
89 |
for user, assistant in history:
|
90 |
+
conversation.extend(
|
91 |
+
[
|
92 |
+
{"role": "user", "content": user},
|
93 |
+
{"role": "assistant", "content": assistant},
|
94 |
+
]
|
95 |
+
)
|
96 |
conversation.append({"role": "user", "content": message})
|
97 |
|
98 |
+
input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(
|
99 |
+
model.device
|
100 |
+
)
|
101 |
+
|
102 |
+
streamer = TextIteratorStreamer(
|
103 |
+
tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True
|
104 |
+
)
|
|
|
|
|
|
|
|
|
|
|
105 |
|
106 |
generate_kwargs = dict(
|
107 |
+
input_ids=input_ids,
|
108 |
streamer=streamer,
|
109 |
max_new_tokens=max_new_tokens,
|
110 |
do_sample=True,
|
111 |
temperature=temperature,
|
112 |
eos_token_id=terminators,
|
113 |
)
|
114 |
+
# This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash.
|
115 |
if temperature == 0:
|
116 |
+
generate_kwargs["do_sample"] = False
|
117 |
+
|
118 |
t = Thread(target=model.generate, kwargs=generate_kwargs)
|
119 |
t.start()
|
120 |
|
121 |
outputs = []
|
122 |
for text in streamer:
|
123 |
outputs.append(text)
|
124 |
+
# print(outputs)
|
125 |
yield "".join(outputs)
|
126 |
+
|
127 |
|
128 |
# Gradio block
|
129 |
+
chatbot = gr.Chatbot(height=450, placeholder=PLACEHOLDER, label="Gradio ChatInterface")
|
130 |
|
131 |
with gr.Blocks(fill_height=True, css=css) as demo:
|
132 |
+
|
133 |
gr.Markdown(DESCRIPTION)
|
|
|
134 |
gr.ChatInterface(
|
135 |
fn=chat_llama3_8b,
|
136 |
chatbot=chatbot,
|
137 |
fill_height=True,
|
138 |
+
additional_inputs_accordion=gr.Accordion(
|
139 |
+
label="⚙️ Parameters", open=False, render=False
|
140 |
+
),
|
141 |
additional_inputs=[
|
142 |
+
gr.Slider(
|
143 |
+
minimum=0,
|
144 |
+
maximum=1,
|
145 |
+
step=0.1,
|
146 |
+
value=0.95,
|
147 |
+
label="Temperature",
|
148 |
+
render=False,
|
149 |
+
),
|
150 |
+
gr.Slider(
|
151 |
+
minimum=128,
|
152 |
+
maximum=4096,
|
153 |
+
step=1,
|
154 |
+
value=512,
|
155 |
+
label="Max new tokens",
|
156 |
+
render=False,
|
157 |
+
),
|
158 |
+
],
|
159 |
examples=[
|
160 |
+
["How to setup a human base on Mars? Give short answer."],
|
161 |
+
["Explain theory of relativity to me like I’m 8 years old."],
|
162 |
+
["What is 9,000 * 9,000?"],
|
163 |
+
["Write a pun-filled happy birthday message to my friend Alex."],
|
164 |
+
["Justify why a penguin might make a good king of the jungle."],
|
165 |
+
],
|
166 |
cache_examples=False,
|
167 |
+
)
|
168 |
+
|
169 |
gr.Markdown(LICENSE)
|
170 |
+
|
171 |
if __name__ == "__main__":
|
172 |
+
demo.launch(share=False)
|
requirements.txt
CHANGED
@@ -2,5 +2,4 @@ triton==2.1.0
|
|
2 |
accelerate
|
3 |
transformers
|
4 |
wheel
|
5 |
-
setuptools
|
6 |
-
pycuda==2023.1
|
|
|
2 |
accelerate
|
3 |
transformers
|
4 |
wheel
|
5 |
+
setuptools
|
|