Spaces:
Runtime error
Runtime error
Chatbot with TinyLlama
Browse files- app.py +120 -0
- imgs/TinyLlama_logo.png +0 -0
- imgs/user_logo.png +0 -0
- requirements.txt +86 -0
app.py
ADDED
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from ctransformers import AutoModelForCausalLM, AutoConfig, Config #import for GGUF/GGML models
|
3 |
+
import datetime
|
4 |
+
|
5 |
+
modelfile="TinyLlama/TinyLlama-1.1B-Chat-v0.6"
|
6 |
+
|
7 |
+
i_temperature = 0.30
|
8 |
+
i_max_new_tokens=1100
|
9 |
+
i_repetitionpenalty = 1.2
|
10 |
+
i_contextlength=12048
|
11 |
+
logfile = 'TinyLlama.1B.txt'
|
12 |
+
|
13 |
+
print("loading model...")
|
14 |
+
|
15 |
+
stt = datetime.datetime.now()
|
16 |
+
conf = AutoConfig(Config(temperature=i_temperature,
|
17 |
+
repetition_penalty=i_repetitionpenalty,
|
18 |
+
batch_size=64,
|
19 |
+
max_new_tokens=i_max_new_tokens,
|
20 |
+
context_length=i_contextlength))
|
21 |
+
llm = AutoModelForCausalLM.from_pretrained(modelfile,
|
22 |
+
model_type="llama",
|
23 |
+
config=conf)
|
24 |
+
dt = datetime.datetime.now() - stt
|
25 |
+
print(f"Model loaded in {dt}")
|
26 |
+
|
27 |
+
def writehistory(text):
|
28 |
+
with open(logfile, 'a', encoding='utf-8') as f:
|
29 |
+
f.write(text)
|
30 |
+
f.write('\n')
|
31 |
+
f.close()
|
32 |
+
|
33 |
+
with gr.Blocks(theme='ParityError/Interstellar') as demo:
|
34 |
+
# TITLE SECTION
|
35 |
+
with gr.Row():
|
36 |
+
with gr.Column(scale=12):
|
37 |
+
gr.HTML("<center>"
|
38 |
+
+ "<h1>🦙 TinyLlama 1.1B 🐋 4K context window</h2></center>")
|
39 |
+
gr.Markdown("""
|
40 |
+
**Currently Running**: [TinyLlama/TinyLlama-1.1B-Chat-v0.6](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v0.6) **Chat History Log File**: *TinyLlama.1B.txt*
|
41 |
+
|
42 |
+
- **Base Model**: TinyLlama/TinyLlama-1.1B-Chat-v0.6, Fine tuned on OpenOrca GPT4 subset for 1 epoch, Using CHATML format.
|
43 |
+
- **License**: Apache 2.0, following the TinyLlama base model.
|
44 |
+
The model output is not censored and the authors do not endorse the opinions in the generated content. Use at your own risk.
|
45 |
+
""")
|
46 |
+
gr.Image(value='imgs/TinyLlama_logo.png', width=70)
|
47 |
+
|
48 |
+
# chat and parameters settings
|
49 |
+
with gr.Row():
|
50 |
+
with gr.Column(scale=4):
|
51 |
+
chatbot = gr.Chatbot(height = 350, show_copy_button=True, avatar_images = ["imgs/user_logo.png","imgs/TinyLlama_logo.png"])
|
52 |
+
with gr.Row():
|
53 |
+
with gr.Column(scale=14):
|
54 |
+
msg = gr.Textbox(show_label=False, placeholder="Enter text", lines=2)
|
55 |
+
submitBtn = gr.Button("\n💬 Send\n", size="lg", variant="primary", min_width=140)
|
56 |
+
|
57 |
+
with gr.Column(min_width=50, scale=1):
|
58 |
+
with gr.Tab(label="Parameter Setting"):
|
59 |
+
gr.Markdown("# Parameters")
|
60 |
+
top_p = gr.Slider(minimum=-0,
|
61 |
+
maximum=1.0,
|
62 |
+
value=0.95,
|
63 |
+
step=0.05,
|
64 |
+
interactive=True,
|
65 |
+
label="Top-p")
|
66 |
+
temperature = gr.Slider(minimum=0.1,
|
67 |
+
maximum=1.0,
|
68 |
+
value=0.30,
|
69 |
+
step=0.01,
|
70 |
+
interactive=True,
|
71 |
+
label="Temperature")
|
72 |
+
max_length_tokens = gr.Slider(minimum=0,
|
73 |
+
maximum=4096,
|
74 |
+
value=1060,
|
75 |
+
step=4,
|
76 |
+
interactive=True,
|
77 |
+
label="Max Generation Tokens")
|
78 |
+
rep_pen = gr.Slider(minimum=0,
|
79 |
+
maximum=5,
|
80 |
+
value=1.2,
|
81 |
+
step=0.05,
|
82 |
+
interactive=True,
|
83 |
+
label="Repetition Penalty")
|
84 |
+
|
85 |
+
clear = gr.Button("🗑️ Clear All Messages", variant='secondary')
|
86 |
+
def user(user_message, history):
|
87 |
+
writehistory(f"USER: {user_message}")
|
88 |
+
return "", history + [[user_message, None]]
|
89 |
+
|
90 |
+
def bot(history, t, p, m, r):
|
91 |
+
SYSTEM_PROMPT = """<|im_start|>system
|
92 |
+
You are a helpful bot. Your answers are clear and concise.
|
93 |
+
<|im_end|>
|
94 |
+
|
95 |
+
"""
|
96 |
+
prompt = f"<|im_start|>system<|im_end|><|im_start|>user\n{history[-1][0]}<|im_end|>\n<|im_start|>assistant\n"
|
97 |
+
print(f"history lenght: {len(history)}")
|
98 |
+
if len(history) == 1:
|
99 |
+
print("this is the first round")
|
100 |
+
else:
|
101 |
+
print("here we should pass more conversations")
|
102 |
+
history[-1][1] = ""
|
103 |
+
for character in llm(prompt,
|
104 |
+
temperature = t,
|
105 |
+
top_p = p,
|
106 |
+
repetition_penalty = r,
|
107 |
+
max_new_tokens=m,
|
108 |
+
stop = ['<|im_end|>'],
|
109 |
+
stream = True):
|
110 |
+
history[-1][1] += character
|
111 |
+
yield history
|
112 |
+
writehistory(f"temperature: {t}, top_p: {p}, maxNewTokens: {m}, repetitionPenalty: {r}\n---\nBOT: {history}\n\n")
|
113 |
+
# Log in the terminal the messages
|
114 |
+
print(f"USER: {history[-1][0]}\n---\ntemperature: {t}, top_p: {p}, maxNewTokens: {m}, repetitionPenalty: {r}\n---\nBOT: {history[-1][1]}\n\n")
|
115 |
+
# Clicking the submitBtn will call the generation with Parameters in the slides
|
116 |
+
submitBtn.click(user, [msg, chatbot], [msg, chatbot], queue=False).then(bot, [chatbot,temperature,top_p,max_length_tokens,rep_pen], chatbot)
|
117 |
+
clear.click(lambda: None, None, chatbot, queue=False)
|
118 |
+
|
119 |
+
demo.queue() # required to yield the streams from the text generation
|
120 |
+
demo.launch(inbrowser=True, share=True)
|
imgs/TinyLlama_logo.png
ADDED
imgs/user_logo.png
ADDED
requirements.txt
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
aiofiles==23.2.1
|
2 |
+
aiohttp==3.9.3
|
3 |
+
aiosignal==1.3.1
|
4 |
+
altair==5.2.0
|
5 |
+
annotated-types==0.6.0
|
6 |
+
anyio==4.3.0
|
7 |
+
async-timeout==4.0.3
|
8 |
+
attrs==23.2.0
|
9 |
+
certifi==2024.2.2
|
10 |
+
charset-normalizer==3.3.2
|
11 |
+
click==8.1.7
|
12 |
+
cmake==3.28.3
|
13 |
+
colorama==0.4.6
|
14 |
+
contourpy==1.2.0
|
15 |
+
cycler==0.12.1
|
16 |
+
exceptiongroup==1.2.0
|
17 |
+
fastapi==0.110.0
|
18 |
+
ffmpy==0.3.2
|
19 |
+
filelock==3.13.1
|
20 |
+
fonttools==4.50.0
|
21 |
+
frozenlist==1.4.1
|
22 |
+
fsspec==2024.3.0
|
23 |
+
gradio==4.21.0
|
24 |
+
gradio_client==0.12.0
|
25 |
+
h11==0.14.0
|
26 |
+
httpcore==1.0.4
|
27 |
+
httpx==0.27.0
|
28 |
+
huggingface-hub==0.21.4
|
29 |
+
idna==3.6
|
30 |
+
importlib_resources==6.3.0
|
31 |
+
Jinja2==3.1.3
|
32 |
+
jsonschema==4.21.1
|
33 |
+
jsonschema-specifications==2023.12.1
|
34 |
+
kiwisolver==1.4.5
|
35 |
+
linkify-it-py==2.0.3
|
36 |
+
lit==18.1.1
|
37 |
+
markdown-it-py==2.2.0
|
38 |
+
MarkupSafe==2.1.5
|
39 |
+
matplotlib==3.8.3
|
40 |
+
mdit-py-plugins==0.3.3
|
41 |
+
mdurl==0.1.2
|
42 |
+
mpmath==1.3.0
|
43 |
+
multidict==6.0.5
|
44 |
+
networkx==3.2.1
|
45 |
+
numpy==1.26.4
|
46 |
+
orjson==3.9.15
|
47 |
+
packaging==24.0
|
48 |
+
pandas==2.2.1
|
49 |
+
pillow==10.2.0
|
50 |
+
pydantic==2.6.4
|
51 |
+
pydantic_core==2.16.3
|
52 |
+
pydub==0.25.1
|
53 |
+
Pygments==2.17.2
|
54 |
+
pyparsing==3.1.2
|
55 |
+
python-dateutil==2.9.0.post0
|
56 |
+
python-multipart==0.0.9
|
57 |
+
pytz==2024.1
|
58 |
+
PyYAML==6.0.1
|
59 |
+
referencing==0.33.0
|
60 |
+
regex==2023.12.25
|
61 |
+
requests==2.31.0
|
62 |
+
rich==13.7.1
|
63 |
+
rpds-py==0.18.0
|
64 |
+
ruff==0.3.3
|
65 |
+
safetensors==0.4.2
|
66 |
+
semantic-version==2.10.0
|
67 |
+
shellingham==1.5.4
|
68 |
+
six==1.16.0
|
69 |
+
sniffio==1.3.1
|
70 |
+
starlette==0.36.3
|
71 |
+
sympy==1.12
|
72 |
+
tokenizers==0.13.3
|
73 |
+
tomlkit==0.12.0
|
74 |
+
toolz==0.12.1
|
75 |
+
torch==2.0.1
|
76 |
+
tqdm==4.66.2
|
77 |
+
transformers==4.31.0
|
78 |
+
triton==2.0.0
|
79 |
+
typer==0.9.0
|
80 |
+
typing_extensions==4.10.0
|
81 |
+
tzdata==2024.1
|
82 |
+
uc-micro-py==1.0.3
|
83 |
+
urllib3==2.2.1
|
84 |
+
uvicorn==0.28.0
|
85 |
+
websockets==11.0.3
|
86 |
+
yarl==1.9.4
|