Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
import gradio as gr
|
2 |
-
|
3 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
4 |
|
5 |
-
|
|
|
6 |
def load_model():
|
7 |
model_id = "Tech-Meld/Hajax_Chat_1.0"
|
8 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
@@ -15,17 +15,20 @@ def get_response(input_text, model, tokenizer):
|
|
15 |
response = tokenizer.decode(outputs[:, inputs.shape[-1]:][0], skip_special_tokens=True)
|
16 |
return response
|
17 |
|
18 |
-
model, tokenizer = load_model()
|
19 |
-
|
20 |
def chat(input_text):
|
21 |
-
|
22 |
-
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
iface = gr.Interface(
|
25 |
chat,
|
26 |
"text",
|
27 |
"text",
|
28 |
-
title="Chat with
|
29 |
description="Type your message and press Enter to chat with the AI.",
|
30 |
)
|
31 |
iface.launch()
|
|
|
1 |
import gradio as gr
|
|
|
2 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
3 |
|
4 |
+
model_cache = {}
|
5 |
+
|
6 |
def load_model():
|
7 |
model_id = "Tech-Meld/Hajax_Chat_1.0"
|
8 |
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
|
|
15 |
response = tokenizer.decode(outputs[:, inputs.shape[-1]:][0], skip_special_tokens=True)
|
16 |
return response
|
17 |
|
|
|
|
|
18 |
def chat(input_text):
|
19 |
+
global model_cache
|
20 |
+
if "model" not in model_cache:
|
21 |
+
model_cache["model"], model_cache["tokenizer"] = load_model()
|
22 |
+
model = model_cache["model"]
|
23 |
+
tokenizer = model_cache["tokenizer"]
|
24 |
+
response = get_response(input_text, model, tokenizer)
|
25 |
+
return response
|
26 |
|
27 |
iface = gr.Interface(
|
28 |
chat,
|
29 |
"text",
|
30 |
"text",
|
31 |
+
title="Chat with AI",
|
32 |
description="Type your message and press Enter to chat with the AI.",
|
33 |
)
|
34 |
iface.launch()
|