fix: change to qwen2
Browse files
app.py
CHANGED
@@ -1,8 +1,17 @@
|
|
1 |
import gradio as gr
|
2 |
from llama_cpp import Llama
|
3 |
|
4 |
-
model = "Qwen/
|
5 |
-
llm = Llama.from_pretrained(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
|
7 |
|
8 |
def respond(
|
|
|
1 |
import gradio as gr
|
2 |
from llama_cpp import Llama
|
3 |
|
4 |
+
model = "Qwen/Qwen2-0.5B-Instruct-GGUF"
|
5 |
+
llm = Llama.from_pretrained(
|
6 |
+
repo_id=model,
|
7 |
+
filename="*q8_0.gguf",
|
8 |
+
verbose=True,
|
9 |
+
use_mmap=False,
|
10 |
+
use_mlock=True,
|
11 |
+
n_threads=2,
|
12 |
+
n_threads_batch=2,
|
13 |
+
n_ctx=32000,
|
14 |
+
)
|
15 |
|
16 |
|
17 |
def respond(
|