Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
@@ -1,51 +1,92 @@
|
|
1 |
import streamlit as st
|
2 |
-
from transformers import
|
3 |
import torch
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
|
5 |
class LlamaDemo:
|
6 |
def __init__(self):
|
7 |
-
self.model_name = "meta-llama/Llama-2-70b-chat"
|
8 |
-
|
9 |
-
self.
|
10 |
|
11 |
@property
|
12 |
-
def
|
13 |
-
if self.
|
14 |
-
self.
|
15 |
-
|
16 |
-
model=self.model_name,
|
17 |
torch_dtype=torch.float16,
|
18 |
device_map="auto",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
trust_remote_code=True
|
20 |
)
|
21 |
-
return self.
|
22 |
|
23 |
-
def generate_response(self, prompt: str,
|
24 |
# Format prompt for Llama 2 chat
|
25 |
formatted_prompt = f"[INST] {prompt} [/INST]"
|
26 |
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
|
|
|
|
|
|
36 |
|
37 |
-
|
38 |
return response.split("[/INST]")[-1].strip()
|
39 |
|
40 |
def main():
|
41 |
st.set_page_config(
|
42 |
-
page_title="Llama 2
|
43 |
page_icon="π¦",
|
44 |
layout="wide"
|
45 |
)
|
46 |
|
47 |
st.title("π¦ Llama 2 Chat Demo")
|
48 |
|
|
|
|
|
|
|
|
|
49 |
# Initialize model
|
50 |
if 'llama' not in st.session_state:
|
51 |
with st.spinner("Loading Llama 2... This might take a few minutes..."):
|
@@ -82,13 +123,6 @@ def main():
|
|
82 |
st.error(f"Error: {str(e)}")
|
83 |
|
84 |
with st.sidebar:
|
85 |
-
st.markdown("""
|
86 |
-
### About
|
87 |
-
This demo uses Llama-2-70B-chat, a large language model from Meta.
|
88 |
-
|
89 |
-
The model runs with automatic device mapping and mixed precision for optimal performance.
|
90 |
-
""")
|
91 |
-
|
92 |
if st.button("Clear Chat History"):
|
93 |
st.session_state.chat_history = []
|
94 |
st.experimental_rerun()
|
|
|
1 |
import streamlit as st
|
2 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
3 |
import torch
|
4 |
+
from huggingface_hub import login
|
5 |
+
import os
|
6 |
+
|
7 |
+
def init_huggingface():
|
8 |
+
"""Initialize Hugging Face authentication either from secrets or user input"""
|
9 |
+
if 'HUGGING_FACE_TOKEN' not in st.session_state:
|
10 |
+
# First try to get from environment variable
|
11 |
+
token = os.getenv('HUGGINGFACE_TOKEN')
|
12 |
+
|
13 |
+
# If not in environment, check streamlit secrets
|
14 |
+
if not token and 'huggingface_token' in st.secrets:
|
15 |
+
token = st.secrets['huggingface_token']
|
16 |
+
|
17 |
+
# If still not found, ask user
|
18 |
+
if not token:
|
19 |
+
token = st.text_input('Enter your Hugging Face token:', type='password')
|
20 |
+
if not token:
|
21 |
+
st.warning('Please enter your Hugging Face token to proceed')
|
22 |
+
st.stop()
|
23 |
+
|
24 |
+
st.session_state['HUGGING_FACE_TOKEN'] = token
|
25 |
+
|
26 |
+
# Login to Hugging Face
|
27 |
+
login(st.session_state['HUGGING_FACE_TOKEN'])
|
28 |
+
return True
|
29 |
|
30 |
class LlamaDemo:
|
31 |
def __init__(self):
|
32 |
+
self.model_name = "meta-llama/Llama-2-70b-chat-hf"
|
33 |
+
self._model = None
|
34 |
+
self._tokenizer = None
|
35 |
|
36 |
@property
|
37 |
+
def model(self):
|
38 |
+
if self._model is None:
|
39 |
+
self._model = AutoModelForCausalLM.from_pretrained(
|
40 |
+
self.model_name,
|
|
|
41 |
torch_dtype=torch.float16,
|
42 |
device_map="auto",
|
43 |
+
trust_remote_code=True,
|
44 |
+
load_in_8bit=True # Para optimizar memoria
|
45 |
+
)
|
46 |
+
return self._model
|
47 |
+
|
48 |
+
@property
|
49 |
+
def tokenizer(self):
|
50 |
+
if self._tokenizer is None:
|
51 |
+
self._tokenizer = AutoTokenizer.from_pretrained(
|
52 |
+
self.model_name,
|
53 |
trust_remote_code=True
|
54 |
)
|
55 |
+
return self._tokenizer
|
56 |
|
57 |
+
def generate_response(self, prompt: str, max_new_tokens: int = 512) -> str:
|
58 |
# Format prompt for Llama 2 chat
|
59 |
formatted_prompt = f"[INST] {prompt} [/INST]"
|
60 |
|
61 |
+
inputs = self.tokenizer(formatted_prompt, return_tensors="pt").to(self.model.device)
|
62 |
+
|
63 |
+
with torch.no_grad():
|
64 |
+
outputs = self.model.generate(
|
65 |
+
**inputs,
|
66 |
+
max_new_tokens=max_new_tokens,
|
67 |
+
num_return_sequences=1,
|
68 |
+
temperature=0.7,
|
69 |
+
do_sample=True,
|
70 |
+
top_p=0.9,
|
71 |
+
pad_token_id=self.tokenizer.eos_token_id
|
72 |
+
)
|
73 |
|
74 |
+
response = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
|
75 |
return response.split("[/INST]")[-1].strip()
|
76 |
|
77 |
def main():
|
78 |
st.set_page_config(
|
79 |
+
page_title="Llama 2 Demo",
|
80 |
page_icon="π¦",
|
81 |
layout="wide"
|
82 |
)
|
83 |
|
84 |
st.title("π¦ Llama 2 Chat Demo")
|
85 |
|
86 |
+
# Initialize Hugging Face authentication
|
87 |
+
if init_huggingface():
|
88 |
+
st.success("Successfully authenticated with Hugging Face!")
|
89 |
+
|
90 |
# Initialize model
|
91 |
if 'llama' not in st.session_state:
|
92 |
with st.spinner("Loading Llama 2... This might take a few minutes..."):
|
|
|
123 |
st.error(f"Error: {str(e)}")
|
124 |
|
125 |
with st.sidebar:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
if st.button("Clear Chat History"):
|
127 |
st.session_state.chat_history = []
|
128 |
st.experimental_rerun()
|