desert
commited on
Commit
•
f613acc
1
Parent(s):
77156ce
update with my model
Browse files
app.py
CHANGED
@@ -1,12 +1,25 @@
|
|
1 |
import gradio as gr
|
2 |
-
from
|
|
|
3 |
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
|
|
|
|
|
10 |
def respond(
|
11 |
message,
|
12 |
history: list[tuple[str, str]],
|
@@ -15,34 +28,44 @@ def respond(
|
|
15 |
temperature,
|
16 |
top_p,
|
17 |
):
|
|
|
18 |
messages = [{"role": "system", "content": system_message}]
|
19 |
|
|
|
20 |
for val in history:
|
21 |
if val[0]:
|
22 |
messages.append({"role": "user", "content": val[0]})
|
23 |
if val[1]:
|
24 |
messages.append({"role": "assistant", "content": val[1]})
|
25 |
|
|
|
26 |
messages.append({"role": "user", "content": message})
|
27 |
|
28 |
-
|
29 |
-
|
30 |
-
for message in client.chat_completion(
|
31 |
messages,
|
32 |
-
|
33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
temperature=temperature,
|
35 |
top_p=top_p,
|
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 |
respond,
|
48 |
additional_inputs=[
|
@@ -59,6 +82,6 @@ demo = gr.ChatInterface(
|
|
59 |
],
|
60 |
)
|
61 |
|
62 |
-
|
63 |
if __name__ == "__main__":
|
64 |
demo.launch()
|
|
|
|
1 |
import gradio as gr
|
2 |
+
from unsloth import FastLanguageModel
|
3 |
+
import torch
|
4 |
|
5 |
+
# Load your model and tokenizer (make sure to adjust the path to where your model is stored)
|
6 |
+
max_seq_length = 2048 # Adjust as necessary
|
7 |
+
load_in_4bit = True # Enable 4-bit quantization for reduced memory usage
|
8 |
+
model_path = "/content/drive/My Drive/llama_lora_model_1" # Path to your custom model
|
9 |
|
10 |
+
# Load the model and tokenizer
|
11 |
+
model, tokenizer = FastLanguageModel.from_pretrained(
|
12 |
+
model_name=model_path,
|
13 |
+
max_seq_length=max_seq_length,
|
14 |
+
load_in_4bit=load_in_4bit,
|
15 |
+
)
|
16 |
+
|
17 |
+
# Move model to GPU if available
|
18 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
19 |
+
model = model.to(device)
|
20 |
|
21 |
+
|
22 |
+
# Respond function
|
23 |
def respond(
|
24 |
message,
|
25 |
history: list[tuple[str, str]],
|
|
|
28 |
temperature,
|
29 |
top_p,
|
30 |
):
|
31 |
+
# Prepare the system message
|
32 |
messages = [{"role": "system", "content": system_message}]
|
33 |
|
34 |
+
# Add history to the messages
|
35 |
for val in history:
|
36 |
if val[0]:
|
37 |
messages.append({"role": "user", "content": val[0]})
|
38 |
if val[1]:
|
39 |
messages.append({"role": "assistant", "content": val[1]})
|
40 |
|
41 |
+
# Add the current message from the user
|
42 |
messages.append({"role": "user", "content": message})
|
43 |
|
44 |
+
# Prepare the inputs for the model
|
45 |
+
inputs = tokenizer.apply_chat_template(
|
|
|
46 |
messages,
|
47 |
+
tokenize=True,
|
48 |
+
add_generation_prompt=True,
|
49 |
+
return_tensors="pt",
|
50 |
+
).to(device)
|
51 |
+
|
52 |
+
# Generate the response using your model
|
53 |
+
outputs = model.generate(
|
54 |
+
input_ids=inputs["input_ids"],
|
55 |
+
max_new_tokens=max_tokens,
|
56 |
temperature=temperature,
|
57 |
top_p=top_p,
|
58 |
+
use_cache=True,
|
59 |
+
)
|
60 |
+
|
61 |
+
# Decode the generated output
|
62 |
+
response = tokenizer.batch_decode(outputs, skip_special_tokens=True)
|
63 |
|
64 |
+
# Return the response
|
65 |
+
return response[0]
|
66 |
|
67 |
|
68 |
+
# Gradio interface setup
|
|
|
|
|
69 |
demo = gr.ChatInterface(
|
70 |
respond,
|
71 |
additional_inputs=[
|
|
|
82 |
],
|
83 |
)
|
84 |
|
|
|
85 |
if __name__ == "__main__":
|
86 |
demo.launch()
|
87 |
+
|