Spaces:
Runtime error
Runtime error
init
Browse files- llava/serve/gradio_web_server.py +10 -2
- pyproject.toml +1 -1
llava/serve/gradio_web_server.py
CHANGED
@@ -31,6 +31,7 @@ from PIL import Image
|
|
31 |
import io
|
32 |
import base64
|
33 |
|
|
|
34 |
def resize_image(image, default_width=(1344, 896), request_width=None):
|
35 |
# 如果 request 中指定了 width,则使用传入的值
|
36 |
if request_width:
|
@@ -38,7 +39,7 @@ def resize_image(image, default_width=(1344, 896), request_width=None):
|
|
38 |
|
39 |
original_width, original_height = image.size
|
40 |
|
41 |
-
print(original_width,original_height)
|
42 |
|
43 |
# 根据宽高比决定 resize 逻辑
|
44 |
if original_width >= original_height:
|
@@ -54,7 +55,14 @@ def resize_image(image, default_width=(1344, 896), request_width=None):
|
|
54 |
|
55 |
# 调整图像大小
|
56 |
resized_image = image.resize((new_width, new_height))
|
57 |
-
print(new_width,new_height)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
58 |
return resized_image
|
59 |
|
60 |
|
|
|
31 |
import io
|
32 |
import base64
|
33 |
|
34 |
+
|
35 |
def resize_image(image, default_width=(1344, 896), request_width=None):
|
36 |
# 如果 request 中指定了 width,则使用传入的值
|
37 |
if request_width:
|
|
|
39 |
|
40 |
original_width, original_height = image.size
|
41 |
|
42 |
+
print("Original size:", original_width, original_height)
|
43 |
|
44 |
# 根据宽高比决定 resize 逻辑
|
45 |
if original_width >= original_height:
|
|
|
55 |
|
56 |
# 调整图像大小
|
57 |
resized_image = image.resize((new_width, new_height))
|
58 |
+
print("After initial resize:", new_width, new_height)
|
59 |
+
|
60 |
+
# 如果高度仍然超过 2016,则将图片固定调整为 896x2016
|
61 |
+
if new_height > 2016:
|
62 |
+
new_width, new_height = 672, 2016
|
63 |
+
resized_image = resized_image.resize((new_width, new_height))
|
64 |
+
print("Adjusted to fixed size:", new_width, new_height)
|
65 |
+
|
66 |
return resized_image
|
67 |
|
68 |
|
pyproject.toml
CHANGED
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
|
|
4 |
|
5 |
[project]
|
6 |
name = "uground_demo_test"
|
7 |
-
version = "3.
|
8 |
description = "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents"
|
9 |
readme = "README.md"
|
10 |
requires-python = ">=3.8"
|
|
|
4 |
|
5 |
[project]
|
6 |
name = "uground_demo_test"
|
7 |
+
version = "3.6"
|
8 |
description = "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents"
|
9 |
readme = "README.md"
|
10 |
requires-python = ">=3.8"
|