Translate Space interface to English

app.py

@@ -62,7 +62,7 @@ def segment_street_scene(
 ):
     """Run semantic segmentation and return visual, tabular, and raw outputs."""
     if image is None:
-        raise gr.Error("请先上传一张街景图片。")
+        raise gr.Error("Please upload a street-scene image first.")
 
     started_at = time.perf_counter()
     prepared_image = resize_for_output(image)
@@ -83,7 +83,7 @@ def segment_street_scene(
         class_map = class_map_tensor.cpu().numpy().astype(np.uint8)
     except Exception as exc:
         raise gr.Error(
-            f"分割模型运行失败：{type(exc).__name__}: {exc}"
+            f"Segmentation failed: {type(exc).__name__}: {exc}"
         ) from exc
 
     overlay, color_mask = render_segmentation(
@@ -109,8 +109,8 @@ def segment_street_scene(
     elapsed = time.perf_counter() - started_at
     visible_classes = len(np.unique(class_map))
     status = (
-        f"完成 · {prepared_image.width}×{prepared_image.height} · "
-        f"{visible_classes} 个街景类别 · {elapsed:.1f}s · "
+        f"Done · {prepared_image.width}×{prepared_image.height} · "
+        f"{visible_classes} street-scene classes · {elapsed:.1f}s · "
         f"device={device.type}"
     )
 
@@ -135,8 +135,8 @@ with gr.Blocks(title="Street Scene Segmentation", theme=gr.themes.Soft(), css=CS
     gr.Markdown(
         """
         <div class="hero">
-          <h1>🚦 街景语义分割</h1>
-          <p>上传道路或城市街景，自动识别道路、人行道、建筑、植被、车辆、行人等 Cityscapes 类别。</p>
+          <h1>🚦 Street Scene Semantic Segmentation</h1>
+          <p>Upload a road or urban street image to identify Cityscapes classes such as roads, sidewalks, buildings, vegetation, vehicles, and people.</p>
           <p class="muted">SegFormer-B0 · 19 street-scene classes · CPU/GPU adaptive</p>
         </div>
         """
@@ -146,53 +146,54 @@ with gr.Blocks(title="Street Scene Segmentation", theme=gr.themes.Soft(), css=CS
         with gr.Column(scale=5):
             image_input = gr.Image(
                 type="pil",
-                label="上传街景图片",
+                label="Upload a street-scene image",
                 height=470,
                 sources=["upload", "clipboard", "webcam"],
             )
-            with gr.Accordion("显示设置", open=False):
+            with gr.Accordion("Display settings", open=False):
                 opacity_input = gr.Slider(
                     0.15,
                     0.85,
                     value=0.55,
                     step=0.05,
-                    label="叠加透明度",
+                    label="Overlay opacity",
                 )
                 min_share_input = gr.Slider(
                     0.0,
                     5.0,
                     value=0.1,
                     step=0.1,
-                    label="类别表最小面积占比 (%)",
+                    label="Minimum class area shown in table (%)",
                 )
             with gr.Row():
-                run_button = gr.Button("开始分割", variant="primary", size="lg")
-                clear_button = gr.ClearButton(value="清空", components=[image_input])
+                run_button = gr.Button("Run segmentation", variant="primary", size="lg")
+                clear_button = gr.ClearButton(value="Clear", components=[image_input])
             status_output = gr.Markdown()
 
         with gr.Column(scale=7):
             with gr.Tabs():
-                with gr.Tab("叠加结果"):
-                    overlay_output = gr.Image(label="语义分割叠加图", height=470)
-                with gr.Tab("彩色掩码"):
-                    mask_output = gr.Image(label="Cityscapes 彩色掩码", height=470)
+                with gr.Tab("Overlay"):
+                    overlay_output = gr.Image(label="Segmentation overlay", height=470)
+                with gr.Tab("Color mask"):
+                    mask_output = gr.Image(label="Cityscapes color mask", height=470)
 
     table_output = gr.Dataframe(
-        headers=["Class ID", "类别", "像素数", "面积占比 (%)", "颜色"],
+        headers=["Class ID", "Class", "Pixels", "Area share (%)", "Color"],
         datatype=["number", "str", "number", "number", "str"],
-        label="检测到的街景类别",
+        label="Detected street-scene classes",
         interactive=False,
         wrap=True,
     )
     files_output = gr.File(
-        label="下载结果（叠加图、彩色掩码、原始 Class ID、CSV）",
+        label="Download results (overlay, color mask, class IDs, and CSV)",
         file_count="multiple",
     )
 
     gr.Markdown(
         """
-        > 首次运行会从 Hugging Face 下载模型，之后复用容器缓存。超大图片会等比例缩放到最长边 2048px，
-        > 以避免免费 Space 内存不足。该模型面向 Cityscapes 街景，不适合医学、遥感或室内场景。
+        > The model is downloaded from Hugging Face on the first run and then reused from the container cache.
+        > Images are resized to a maximum side length of 2048 px to prevent memory issues.
+        > This Cityscapes model is intended for street scenes, not medical, remote-sensing, or indoor imagery.
         """
     )