Spaces:
Running
on
Zero
Running
on
Zero
Create app.py
Browse files
app.py
CHANGED
@@ -1,64 +1,62 @@
|
|
1 |
-
import gradio as gr
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
-
|
5 |
-
|
6 |
-
"""
|
7 |
-
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
|
8 |
|
|
|
|
|
9 |
|
10 |
-
def
|
11 |
-
|
12 |
-
history: list[tuple[str, str]],
|
13 |
-
system_message,
|
14 |
-
max_tokens,
|
15 |
-
temperature,
|
16 |
-
top_p,
|
17 |
-
):
|
18 |
-
messages = [{"role": "system", "content": system_message}]
|
19 |
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
if val[1]:
|
24 |
-
messages.append({"role": "assistant", "content": val[1]})
|
25 |
|
26 |
-
|
27 |
|
28 |
-
|
29 |
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
temperature=
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
|
39 |
-
|
40 |
-
|
|
|
41 |
|
|
|
42 |
|
43 |
-
"""
|
44 |
-
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
|
45 |
-
"""
|
46 |
demo = gr.ChatInterface(
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
step=0.05,
|
57 |
-
label="Top-p (nucleus sampling)",
|
58 |
-
),
|
59 |
],
|
|
|
|
|
|
|
60 |
)
|
61 |
|
62 |
-
|
63 |
if __name__ == "__main__":
|
64 |
-
demo.launch()
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import torch
|
3 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
4 |
+
from peft import PeftModel
|
5 |
+
|
6 |
+
base_model = AutoModelForCausalLM.from_pretrained(
|
7 |
+
"unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit",
|
8 |
+
torch_dtype=torch.float16,
|
9 |
+
device_map="auto",
|
10 |
+
load_in_4bit=True
|
11 |
+
)
|
12 |
|
13 |
+
#tokenizer
|
14 |
+
tokenizer = AutoTokenizer.from_pretrained("unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit")
|
|
|
|
|
15 |
|
16 |
+
#LoRA adaptors
|
17 |
+
model = PeftModel.from_pretrained(base_model, "rezaenayati/RezAi-Model")
|
18 |
|
19 |
+
def chat_with_rezAi(messages, history):
|
20 |
+
conversation = "<|start_header_id|>system<|end_header_id|>\nYou are Reza Enayati, a Computer Science student and entrepreneur from Los Angeles, who is eager to work as a software engineer or machine learning engineer. Answer these questions as if you are in an interview.<|eot_id|>"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
22 |
+
for user_msg, assistant_msg in history:
|
23 |
+
conversation += f"<|start_header_id|>user<|end_header_id|>\n{user_msg}<|eot_id|>"
|
24 |
+
conversation += f"<|start_header_id|>assistant<|end_header_id|>\n{assistant_msg}<|eot_id|>"
|
|
|
|
|
25 |
|
26 |
+
conversation += f"<|start_header_id|>user<|end_header_id|>\n{message}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n"
|
27 |
|
28 |
+
inputs = tokenizer([conversation], return_tensors="pt")
|
29 |
|
30 |
+
with torch.no_grad():
|
31 |
+
outputs = model.generate(
|
32 |
+
**inputs,
|
33 |
+
max_new_tokens=128,
|
34 |
+
temperature=0.5,
|
35 |
+
do_sample=True,
|
36 |
+
pad_token_id=tokenizer.eos_token_id
|
37 |
+
)
|
38 |
|
39 |
+
#get response
|
40 |
+
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
41 |
+
new_response = response.split("<|start_header_id|>assistant<|end_header_id|>")[-1].strip()
|
42 |
|
43 |
+
return new_response
|
44 |
|
|
|
|
|
|
|
45 |
demo = gr.ChatInterface(
|
46 |
+
fn=chat_with_rezAi,
|
47 |
+
title="💬 Chat with RezAI",
|
48 |
+
description="Hi! I'm RezAI. Ask me about his technical background, projects, or experience!",
|
49 |
+
examples=[
|
50 |
+
"Tell me about your background",
|
51 |
+
"What programming languages do you know?",
|
52 |
+
"Walk me through your Pizza Guys project",
|
53 |
+
"What's your experience with machine learning?",
|
54 |
+
"How did you get into computer science?"
|
|
|
|
|
|
|
55 |
],
|
56 |
+
retry_btn=None,
|
57 |
+
undo_btn="Delete Previous",
|
58 |
+
clear_btn="Clear Chat",
|
59 |
)
|
60 |
|
|
|
61 |
if __name__ == "__main__":
|
62 |
+
demo.launch()
|