Spaces:
Runtime error
Runtime error
Mikhil-jivus
commited on
Commit
•
b17ecc2
1
Parent(s):
ae4333a
Update app.py
Browse files
app.py
CHANGED
@@ -18,6 +18,16 @@ model = AutoModelForCausalLM.from_pretrained(
|
|
18 |
device_map="auto" # Automatically use available GPU/CPU efficiently
|
19 |
)
|
20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
def respond(
|
22 |
message,
|
23 |
history: list[tuple[str, str]],
|
@@ -26,18 +36,17 @@ def respond(
|
|
26 |
temperature,
|
27 |
top_p,
|
28 |
):
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
if val[1]:
|
35 |
-
messages.append({"role": "assistant", "content": val[1]})
|
36 |
|
37 |
-
|
|
|
|
|
38 |
|
39 |
# Tokenize the input messages
|
40 |
-
input_text = system_message + " ".join([f"{msg['role']}: {msg['content']}" for msg in messages])
|
41 |
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
42 |
|
43 |
# Move input_ids to the GPU
|
@@ -60,15 +69,19 @@ def respond(
|
|
60 |
# Decode the response
|
61 |
response = tokenizer.decode(chat_history_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)
|
62 |
|
63 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
64 |
|
65 |
-
|
66 |
-
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
|
67 |
-
"""
|
68 |
demo = gr.ChatInterface(
|
69 |
respond,
|
70 |
additional_inputs=[
|
71 |
-
gr.Textbox(value="You are a friendly
|
72 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
|
73 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
74 |
gr.Slider(
|
@@ -82,4 +95,4 @@ demo = gr.ChatInterface(
|
|
82 |
)
|
83 |
|
84 |
if __name__ == "__main__":
|
85 |
-
demo.launch(share=True)
|
|
|
18 |
device_map="auto" # Automatically use available GPU/CPU efficiently
|
19 |
)
|
20 |
|
21 |
+
# Define a function to clean up any repeated segments in the generated response
|
22 |
+
def clean_response(response, history):
|
23 |
+
# Check for repetition in the response and remove it
|
24 |
+
if len(history) > 0:
|
25 |
+
last_user_message, last_bot_response = history[-1]
|
26 |
+
if last_bot_response in response:
|
27 |
+
response = response.replace(last_bot_response, "").strip()
|
28 |
+
|
29 |
+
return response
|
30 |
+
|
31 |
def respond(
|
32 |
message,
|
33 |
history: list[tuple[str, str]],
|
|
|
36 |
temperature,
|
37 |
top_p,
|
38 |
):
|
39 |
+
# Add system prompt only once at the beginning of the conversation
|
40 |
+
if len(history) == 0:
|
41 |
+
input_text = f"system: {system_message}\nuser: {message}\n"
|
42 |
+
else:
|
43 |
+
input_text = f"user: {message}\n"
|
|
|
|
|
44 |
|
45 |
+
# Append previous conversation history to the input text
|
46 |
+
for user_msg, bot_msg in history:
|
47 |
+
input_text += f"user: {user_msg}\nassistant: {bot_msg}\n"
|
48 |
|
49 |
# Tokenize the input messages
|
|
|
50 |
input_ids = tokenizer.encode(input_text, return_tensors="pt")
|
51 |
|
52 |
# Move input_ids to the GPU
|
|
|
69 |
# Decode the response
|
70 |
response = tokenizer.decode(chat_history_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)
|
71 |
|
72 |
+
# Clean the response to remove any repeated or unnecessary text
|
73 |
+
response = clean_response(response, history)
|
74 |
+
|
75 |
+
# Update history with the new user message and bot response
|
76 |
+
history.append((message, response))
|
77 |
+
|
78 |
+
return response
|
79 |
|
80 |
+
# Set up the Gradio app interface
|
|
|
|
|
81 |
demo = gr.ChatInterface(
|
82 |
respond,
|
83 |
additional_inputs=[
|
84 |
+
gr.Textbox(value="You are a helpful and friendly assistant.", label="System message"),
|
85 |
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
|
86 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
87 |
gr.Slider(
|
|
|
95 |
)
|
96 |
|
97 |
if __name__ == "__main__":
|
98 |
+
demo.launch(share=True)
|