Unify companion-repo labels: Project page / Dataset / Benchmark code / Model & Inference code

Header bar, Submit-tab instructions, and About-tab companion list now
all use the same five labels as the project page and the rest of the
suite. Drops the older 📦 Code / Reference model framing.

README.md

@@ -21,13 +21,13 @@ The full thirteen-metric vector is the unit of report. The table is sorted by **
 
 ## Submitting
 
-1. Run the official benchmark on the 157-clip frozen evaluation split: `bash scripts/run_benchmark.sh <your_method>` in the [code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench).
+1. Run the official benchmark on the 157-clip frozen evaluation split: `bash scripts/run_benchmark.sh <your_method>` in the [Benchmark code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench).
 2. Upload the produced `outputs/<method>/eval.json` via the **Submit** tab.
 3. The maintainers review the submission. Approved entries appear on the leaderboard.
 
 ## Companion repos
 
-- 🌐 Project page: https://vitex-bench.github.io/
-- 📦 Code: https://huggingface.co/ViTeX-Bench/ViTeX-Bench
-- 📊 Dataset (387 paired clips, 230 train / 157 frozen test): https://huggingface.co/datasets/ViTeX-Bench/ViTeX-Dataset
-- 🤖 Reference model (ViTeX-14B): https://huggingface.co/ViTeX-Bench/ViTeX-14B
+- 🌐 **Project page:** https://vitex-bench.github.io/
+- 📊 **Dataset:** https://huggingface.co/datasets/ViTeX-Bench/ViTeX-Dataset
+- 🧪 **Benchmark code:** https://huggingface.co/ViTeX-Bench/ViTeX-Bench
+- 🤖 **Model & Inference code** (ViTeX-14B): https://huggingface.co/ViTeX-Bench/ViTeX-14B

app.py

@@ -550,11 +550,10 @@ def admin_act(passphrase: str, idx: int, action: str):
 
 HEADER_MD = (
     "# 🏆 ViTeX-Bench Leaderboard\n\n"
-    "🌐 **Project page: [vitex-bench.github.io](https://vitex-bench.github.io/)** "
-    " ·  "
-    "📦 [Code](https://huggingface.co/ViTeX-Bench/ViTeX-Bench)  ·  "
+    "🌐 [Project page](https://vitex-bench.github.io/)  ·  "
     "📊 [Dataset](https://huggingface.co/datasets/ViTeX-Bench/ViTeX-Dataset)  ·  "
-    "🤖 [ViTeX-14B](https://huggingface.co/ViTeX-Bench/ViTeX-14B)\n\n"
+    "🧪 [Benchmark code](https://huggingface.co/ViTeX-Bench/ViTeX-Bench)  ·  "
+    "🤖 [Model & Inference code](https://huggingface.co/ViTeX-Bench/ViTeX-14B)\n\n"
     "Public ranking for **video scene text editing** under the 13-metric, three-axis "
     "[ViTeX-Bench](https://huggingface.co/ViTeX-Bench/ViTeX-Bench) protocol. "
     "The full thirteen-metric vector is the unit of report; the table is sorted by "
@@ -562,7 +561,7 @@ HEADER_MD = (
     "text-correctness primitives. TextScore is a single-axis sort key by design — "
     "no cross-axis aggregate is computed, because no axis substitutes for another.\n\n"
     "**To submit:** run `bash scripts/run_benchmark.sh <your_method>` in the "
-    "[code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench), then upload "
+    "[Benchmark code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench), then upload "
     "the resulting `eval.json` in the **Submit** tab. Submissions are reviewed "
     "before appearing on the leaderboard."
 )
@@ -580,7 +579,7 @@ with gr.Blocks(title="ViTeX-Bench Leaderboard", theme=gr.themes.Soft()) as demo:
         gr.Markdown(
             "Upload your `eval.json` produced by "
             "`bash scripts/run_benchmark.sh <method>` in the "
-            "[code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench). "
+            "[Benchmark code repo](https://huggingface.co/ViTeX-Bench/ViTeX-Bench). "
             "All 13 metrics + TextScore are read directly from the JSON; "
             "you only need to fill the method metadata below."
         )
@@ -661,9 +660,9 @@ with gr.Blocks(title="ViTeX-Bench Leaderboard", theme=gr.themes.Soft()) as demo:
             "ten primitives still appear next to TextScore on the table.\n\n"
             "### Companion repos\n\n"
             "- 🌐 Project page: https://vitex-bench.github.io/\n"
-            "- 📦 Code: https://huggingface.co/ViTeX-Bench/ViTeX-Bench\n"
             "- 📊 Dataset: https://huggingface.co/datasets/ViTeX-Bench/ViTeX-Dataset\n"
-            "- 🤖 Reference model: https://huggingface.co/ViTeX-Bench/ViTeX-14B\n\n"
+            "- 🧪 Benchmark code: https://huggingface.co/ViTeX-Bench/ViTeX-Bench\n"
+            "- 🤖 Model & Inference code: https://huggingface.co/ViTeX-Bench/ViTeX-14B\n\n"
             "Paper under double-blind review at NeurIPS 2026 D&B Track. "
             "Author list and DOI updated after deanonymization."
         )