update

app.py

@@ -221,13 +221,13 @@ with demo:
                             elem_id="column-select",
                             interactive=True,
                         )
-                    with gr.Row():
-                        hide_models = gr.CheckboxGroup(
-                            label="Hide models",
-                            choices = ["Private or deleted", "Contains a merge/moerge", "Flagged", "MoE"],
-                            value=[],
-                            interactive=True
-                        )
+                    #with gr.Row():
+                    #    hide_models = gr.CheckboxGroup(
+                    #        label="Hide models",
+                    #        choices = ["Private or deleted", "Contains a merge/moerge", "Flagged", "MoE"],
+                    #        value=[],
+                    #        interactive=True
+                    #    )
                 with gr.Column(min_width=320):
                     #with gr.Box(elem_id="box-filter"):
                     filter_columns_type = gr.CheckboxGroup(
@@ -349,13 +349,13 @@ with demo:
                             elem_id="column-select",
                             interactive=True,
                         )
-                    with gr.Row():
-                        hide_models = gr.CheckboxGroup(
-                            label="Hide models",
-                            choices = ["Private or deleted", "Contains a merge/moerge", "Flagged", "MoE"],
-                            value=[],
-                            interactive=True
-                        )
+                    #with gr.Row():
+                    #    hide_models = gr.CheckboxGroup(
+                    #        label="Hide models",
+                    #        choices = ["Private or deleted", "Contains a merge/moerge", "Flagged", "MoE"],
+                    #        value=[],
+                    #        interactive=True
+                    #    )
                 with gr.Column(min_width=320):
                     #with gr.Box(elem_id="box-filter"):
                     filter_columns_type = gr.CheckboxGroup(
@@ -450,6 +450,7 @@ with demo:
                     leaderboard_table,
                     queue=True,
                 )
+            gr.Markdown(EVALUATION_METRIC_TEXT, elem_classes="markdown-text")
         with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=2):
             gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
 

src/about.py

@@ -25,23 +25,23 @@ class Tasks(Enum):
 
 class Quotas(Enum):
     Perception = Task("Perception", "overall", "Perception")
-    Perception_VisualGrounding = Task("Perception", "Visual Grounding", "Perception_VisualGrounding")
-    Perception_Counting = Task("Perception", "Counting", "Perception_Counting")
-    Perception_StateActivity_Understanding = Task("Perception", "State & Activity Understanding", "Perception_StateActivity-Understanding")
+    Perception_VisualGrounding = Task("Perception", "Visual Grounding", "P_VG")
+    Perception_Counting = Task("Perception", "Counting", "P_C")
+    Perception_StateActivity_Understanding = Task("Perception", "State & Activity Understanding", "P_SAU")
     SpatialReasoning = Task("SpatialReasoning", "overall", "SpatialReasoning")
-    SpatialReasoning_Dynamic = Task("SpatialReasoning", "Dynamic", "SpatialReasoning_Dynamic")
-    SpatialReasoning_Rd = Task("SpatialReasoning", "Relative direction", "SpatialReasoning_Relative-direction")
-    SpatialReasoning_mmm = Task("SpatialReasoning", "Multi-view matching", "SpatialReasoning_Multi-view-matching")
-    SpatialReasoning_red = Task("SpatialReasoning", "Relative distance", "SpatialReasoning_Relative-distance")
-    SpatialReasoning_ds = Task("SpatialReasoning", "Depth estimation", "SpatialReasoning_Depth-estimation")
-    SpatialReasoning_rs = Task("SpatialReasoning", "Relative shape", "SpatialReasoning_Relative-shape")
-    SpatialReasoning_se = Task("SpatialReasoning", "Size estimation", "SpatialReasoning_Size-estimation")
+    SpatialReasoning_Dynamic = Task("SpatialReasoning", "Dynamic", "SR_D")
+    SpatialReasoning_Rd = Task("SpatialReasoning", "Relative direction", "SR_R")
+    SpatialReasoning_mmm = Task("SpatialReasoning", "Multi-view matching", "SR_Mm")
+    SpatialReasoning_red = Task("SpatialReasoning", "Relative distance", "SR_Rd")
+    SpatialReasoning_ds = Task("SpatialReasoning", "Depth estimation", "SR_De")
+    SpatialReasoning_rs = Task("SpatialReasoning", "Relative shape", "SR_Rs")
+    SpatialReasoning_se = Task("SpatialReasoning", "Size estimation", "SR_Se")
     Prediction = Task("Prediction", "overall", "Prediction")
     Prediction_T = Task("Prediction", "Trajectory", "Prediction_Trajectory")
     Prediction_F = Task("Prediction", "Future prediction", "Prediction_Future-prediction")
     Planning = Task("Planning", "overall", "Planning")
-    Planning_G = Task("Planning", "Goal Decomposition", "Planning_Goal-Decomposition")
-    Planning_N = Task("Planning", "Navigation", "Planning_Navigation")
+    Planning_G = Task("Planning", "Goal Decomposition", "P_GD")
+    Planning_N = Task("Planning", "Navigation", "P_N")
 
 NUM_FEWSHOT = 0 # Change with your few shot
 # ---------------------------------------------------
@@ -109,6 +109,30 @@ You can find:
 
 """
 
+
+EVALUATION_METRIC_TEXT = """
+### 评测指标缩写介绍如下：
+### Evaluation Metrics Abbreviations are introduced below:
+Perception
+- Perception_Visual Grounding(P_VG)
+- Perception_Counting(P_C)
+- Perception_State & Activity Understanding
+SpatialReasoning
+- SpatialReasoning_Dynamic(SR_D)
+- SpatialReasoning_Relative direction(SR_Rd)
+- SpatialReasoning_Multi-view matching(SR_Mm)
+- SpatialReasoning_Relative distance(SR_Rd)
+- SpatialReasoning_Depth estimation(SR_De)
+- SpatialReasoning_Relative shape(SR_Rs)
+- SpatialReasoning_Size estimation(SR_Se)
+Prediction
+- Prediction_Trajectory(P_T)
+- Prediction_Futureprediction(P_Fd)
+Planning
+- Planning_Goal Decomposition(P_GD)
+- Planning_Navigation(P_N)
+"""
+
 EVALUATION_QUEUE_TEXT = """
 ## Evaluation Queue for the FlagEval VLM Leaderboard
 Models added here will be automatically evaluated on the FlagEval cluster.