Update app.py
Browse files
app.py
CHANGED
|
@@ -47,23 +47,119 @@ def generate_caption(image_path, question):
|
|
| 47 |
full_response += token
|
| 48 |
yield full_response
|
| 49 |
|
|
|
|
|
|
|
| 50 |
# 创建Gradio界面
|
| 51 |
-
title="
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
question_input = gr.Textbox(label="输入问题", value="请描述图片内容")
|
| 58 |
-
submit_btn = gr.Button("生成描述")
|
| 59 |
-
output = gr.Textbox(label="描述结果", interactive=False)
|
| 60 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
submit_btn.click(
|
| 62 |
fn=generate_caption,
|
| 63 |
inputs=[image_input, question_input],
|
| 64 |
-
outputs=output
|
|
|
|
| 65 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
|
| 67 |
if __name__ == "__main__":
|
| 68 |
demo.queue(default_concurrency_limit=100)
|
| 69 |
-
demo.launch(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
full_response += token
|
| 48 |
yield full_response
|
| 49 |
|
| 50 |
+
|
| 51 |
+
|
| 52 |
# 创建Gradio界面
|
| 53 |
+
title = "Hunyuan-Vision图生文Demo"
|
| 54 |
+
theme = gr.themes.Soft(
|
| 55 |
+
primary_hue="teal",
|
| 56 |
+
secondary_hue="blue",
|
| 57 |
+
font=[gr.themes.GoogleFont("Noto Sans SC"), "Arial", "sans-serif"]
|
| 58 |
+
)
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
+
with gr.Blocks(title=title, theme=theme) as demo:
|
| 61 |
+
# ================= 头部区域 =================
|
| 62 |
+
gr.Markdown(f"""
|
| 63 |
+
<div style="text-align: center;">
|
| 64 |
+
<h1 style="color: #2E86C1; border-bottom: 3px solid #AED6F1; padding-bottom: 10px;">🖼️ {title}</h1>
|
| 65 |
+
<p style="color: #616A6B;">上传图片并输入问题,体验腾讯混元视觉大模型的图像理解能力</p>
|
| 66 |
+
</div>
|
| 67 |
+
""")
|
| 68 |
+
|
| 69 |
+
# ================= 主体区域 =================
|
| 70 |
+
with gr.Row(variant="panel"):
|
| 71 |
+
# 左侧输入列
|
| 72 |
+
with gr.Column(scale=3):
|
| 73 |
+
with gr.Group(label="输入区域"):
|
| 74 |
+
image_input = gr.Image(
|
| 75 |
+
type="filepath",
|
| 76 |
+
label="上传图片",
|
| 77 |
+
height=400,
|
| 78 |
+
show_download_button=False,
|
| 79 |
+
elem_classes="preview-box"
|
| 80 |
+
)
|
| 81 |
+
question_input = gr.Textbox(
|
| 82 |
+
label="问题描述",
|
| 83 |
+
placeholder="请输入关于图片的问题...",
|
| 84 |
+
value="请详细描述图片中的场景、人物和细节",
|
| 85 |
+
lines=2
|
| 86 |
+
)
|
| 87 |
+
with gr.Row():
|
| 88 |
+
clear_btn = gr.Button("清空", variant="secondary")
|
| 89 |
+
submit_btn = gr.Button("生成描述", variant="primary")
|
| 90 |
+
|
| 91 |
+
# 右侧输出列
|
| 92 |
+
with gr.Column(scale=4):
|
| 93 |
+
with gr.Group(label="生成结果"):
|
| 94 |
+
output = gr.Textbox(
|
| 95 |
+
label="描述内容",
|
| 96 |
+
interactive=False,
|
| 97 |
+
show_copy_button=True,
|
| 98 |
+
lines=12,
|
| 99 |
+
max_lines=20,
|
| 100 |
+
autoscroll=True
|
| 101 |
+
)
|
| 102 |
+
|
| 103 |
+
# ================= 示例区域 =================
|
| 104 |
+
with gr.Accordion("🖼️ 点击查看示例", open=False):
|
| 105 |
+
with gr.Row():
|
| 106 |
+
gr.Examples(
|
| 107 |
+
examples=[
|
| 108 |
+
["tencent.png", "图片中的天气状况如何?"],
|
| 109 |
+
["tencent.png", "描述参会人员的衣着特征"]
|
| 110 |
+
],
|
| 111 |
+
inputs=[image_input, question_input],
|
| 112 |
+
label="快速示例"
|
| 113 |
+
)
|
| 114 |
+
|
| 115 |
+
# ================= 交互逻辑 =================
|
| 116 |
submit_btn.click(
|
| 117 |
fn=generate_caption,
|
| 118 |
inputs=[image_input, question_input],
|
| 119 |
+
outputs=output,
|
| 120 |
+
api_name="generate"
|
| 121 |
)
|
| 122 |
+
|
| 123 |
+
clear_btn.click(
|
| 124 |
+
fn=lambda: [None, "", ""],
|
| 125 |
+
outputs=[image_input, question_input, output],
|
| 126 |
+
queue=False
|
| 127 |
+
)
|
| 128 |
+
|
| 129 |
+
# ================= 自定义样式 =================
|
| 130 |
+
css = """
|
| 131 |
+
.preview-box img {border-radius: 10px; box-shadow: 0 4px 6px rgba(0,0,0,0.1);}
|
| 132 |
+
.preview-box:hover img {transform: scale(1.02);}
|
| 133 |
+
button#generate {transition: all 0.3s ease;}
|
| 134 |
+
"""
|
| 135 |
+
demo.css = css
|
| 136 |
|
| 137 |
if __name__ == "__main__":
|
| 138 |
demo.queue(default_concurrency_limit=100)
|
| 139 |
+
demo.launch(
|
| 140 |
+
server_port=7860,
|
| 141 |
+
show_error=True,
|
| 142 |
+
favicon_path="favicon.ico",
|
| 143 |
+
max_threads=100
|
| 144 |
+
)
|
| 145 |
+
|
| 146 |
+
# # 创建Gradio界面
|
| 147 |
+
# title="Hunyuan-Vision图生文Demo"
|
| 148 |
+
# with gr.Blocks(title=title) as demo:
|
| 149 |
+
# gr.Markdown(f"# 🖼️ {title}")
|
| 150 |
+
# with gr.Row():
|
| 151 |
+
# with gr.Column():
|
| 152 |
+
# image_input = gr.Image(type="filepath", label="上传图片")
|
| 153 |
+
# question_input = gr.Textbox(label="输入问题", value="请描述图片内容")
|
| 154 |
+
# submit_btn = gr.Button("生成描述")
|
| 155 |
+
# output = gr.Textbox(label="描述结果", interactive=False)
|
| 156 |
+
|
| 157 |
+
# submit_btn.click(
|
| 158 |
+
# fn=generate_caption,
|
| 159 |
+
# inputs=[image_input, question_input],
|
| 160 |
+
# outputs=output
|
| 161 |
+
# )
|
| 162 |
+
|
| 163 |
+
# if __name__ == "__main__":
|
| 164 |
+
# demo.queue(default_concurrency_limit=100)
|
| 165 |
+
# demo.launch(max_threads=100)
|