Hugging Face's logo

Spaces:
Urdatorn
/
sphragis-leaderboard
Running

App Files Community
Urdatorn commited on
Commit
9737977
·
1 Parent(s): c79e49f

Build static Sphragis leaderboard

Browse files
Files changed (2) hide show
  1. README.md +26 -6
  2. index.html +81 -48
README.md CHANGED
@@ -1,11 +1,31 @@
1
  ---
2
  title: Sphragis Leaderboard
3
- emoji: 🖼️
4
- colorFrom: yellow
5
- colorTo: red
6
  sdk: static
7
- pinned: false
8
- short_description: 'Leaderboard for grc author attribution '
 
 
 
 
 
9
  ---
10
 
11
- Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  title: Sphragis Leaderboard
3
+ emoji: 🥇
4
+ colorFrom: green
5
+ colorTo: indigo
6
  sdk: static
7
+ pinned: true
8
+ license: apache-2.0
9
+ short_description: Leaderboard for Ancient Greek author attribution
10
+ tags:
11
+ - leaderboard
12
+ - authorship-attribution
13
+ - ancient-greek
14
  ---
15
 
16
+ # Sphragis Leaderboard
17
+
18
+ A serverless Gradio-Lite leaderboard for the nine Sphragis Ancient Greek
19
+ authorship-attribution tasks. It runs entirely in the visitor's browser and
20
+ does not require CPU or GPU Space hardware.
21
+
22
+ Scores are embedded as `results.csv` in `index.html` and are test macro-F1
23
+ percentages. Add one row per reproducible model configuration. If validation
24
+ produces statistically or numerically tied candidates, report every tied test
25
+ result (or a range); never select a winner using the test split.
26
+
27
+ To test locally, serve the repository over HTTP and open the displayed URL:
28
+
29
+ ```bash
30
+ python3 -m http.server 8000
31
+ ```
index.html CHANGED
@@ -1,57 +1,90 @@
1
- <!DOCTYPE html>
2
- <html>
3
- <head>
4
- <meta charset="utf-8">
5
- <meta name="viewport" content="width=device-width, initial-scale=1">
6
- <title>Gradio-Lite: Serverless Gradio Running Entirely in Your Browser</title>
7
- <meta name="description" content="Gradio-Lite: Serverless Gradio Running Entirely in Your Browser">
8
-
9
- <script type="module" crossorigin src="https://cdn.jsdelivr.net/npm/@gradio/lite/dist/lite.js"></script>
10
- <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@gradio/lite/dist/lite.css" />
11
-
12
- <style>
13
- html, body {
14
- margin: 0;
15
- padding: 0;
16
- height: 100%;
17
- }
18
- </style>
19
- </head>
20
- <body>
21
- <gradio-lite>
22
- <gradio-file name="app.py" entrypoint>
23
  import gradio as gr
24
 
25
- from filters import as_gray
26
 
27
- def process(input_image):
28
- output_image = as_gray(input_image)
29
- return output_image
 
 
 
30
 
31
- demo = gr.Interface(
32
- process,
33
- "image",
34
- "image",
35
- examples=["lion.jpg", "logo.png"],
36
- )
37
 
38
- demo.launch()
39
- </gradio-file>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
40
 
41
- <gradio-file name="filters.py">
42
- from skimage.color import rgb2gray
43
 
44
- def as_gray(image):
45
- return rgb2gray(image)
46
- </gradio-file>
47
 
48
- <gradio-file name="lion.jpg" url="https://raw.githubusercontent.com/gradio-app/gradio/main/gradio/test_data/lion.jpg" />
49
- <gradio-file name="logo.png" url="https://raw.githubusercontent.com/gradio-app/gradio/main/guides/assets/logo.png" />
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
50
 
51
- <gradio-requirements>
52
- # Same syntax as requirements.txt
53
- scikit-image
54
- </gradio-requirements>
55
- </gradio-lite>
56
- </body>
57
- </html>
 
1
+ <!doctype html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="utf-8" />
5
+ <meta name="viewport" content="width=device-width, initial-scale=1" />
6
+ <title>Sphragis benchmark leaderboard</title>
7
+ <meta name="description" content="Leaderboard for the Sphragis Ancient Greek authorship-attribution benchmark." />
8
+ <script type="module" crossorigin src="https://cdn.jsdelivr.net/npm/@gradio/lite/dist/lite.js"></script>
9
+ <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@gradio/lite/dist/lite.css" />
10
+ <style>
11
+ html, body { margin: 0; min-height: 100%; }
12
+ gradio-lite { display: block; min-height: 100vh; }
13
+ </style>
14
+ </head>
15
+ <body>
16
+ <gradio-lite>
17
+ <gradio-file name="app.py" entrypoint>
18
+ import csv
19
+
 
 
 
20
  import gradio as gr
21
 
 
22
 
23
+ def load_results():
24
+ with open("results.csv", encoding="utf-8", newline="") as source:
25
+ rows = list(csv.reader(source))
26
+ if len(rows) &lt; 2:
27
+ raise ValueError("No leaderboard rows found in results.csv")
28
+ return rows[0], rows[1:]
29
 
 
 
 
 
 
 
30
 
31
+ INTRODUCTION = """
32
+ # Sphragis benchmark leaderboard
33
+
34
+ Author attribution for Ancient Greek prose, sentence-aligned verse, and
35
+ metrical lines. Every score is **test macro-F1 (%)**. The suffix is the number
36
+ of sentences—or lines for `verse_metre`—in each evaluation example.
37
+
38
+ Models are compared on the fixed, provenance-preserving Sphragis splits.
39
+ Validation selects checkpoints and hyperparameters; test results are reported
40
+ without using test performance to break validation ties.
41
+ """
42
+
43
+ METHODOLOGY = """
44
+ ### Notes
45
+
46
+ † Four GreBerta heads tied at validation macro-F1 1.0000 for
47
+ `verse_sentence_100`. Their test macro-F1 scores were **92.82, 100.00, 100.00,
48
+ and 92.51**, so the table reports the full **92.51–100.00** range rather than
49
+ choosing the best test result.
50
+
51
+ ‡ Four GreBerta heads tied at validation macro-F1 1.0000 for
52
+ `verse_metre_100`; all four obtained test macro-F1 **96.10**.
53
 
54
+ The variable-chunk GreBerta `_100` repair experiment is marked exploratory
55
+ because its design followed inspection of an earlier `_100` test failure.
56
 
57
+ ### Benchmark resources
 
 
58
 
59
+ - [Dataset](https://huggingface.co/datasets/Urdatorn/sphragis)
60
+ - [Model implementations](https://github.com/Urdatorn/sphragis_models)
61
+ """
62
+
63
+
64
+ headers, values = load_results()
65
+
66
+ with gr.Blocks(title="Sphragis Leaderboard") as demo:
67
+ gr.Markdown(INTRODUCTION)
68
+ gr.Dataframe(
69
+ headers=headers,
70
+ value=values,
71
+ datatype=["str"] * len(headers),
72
+ interactive=False,
73
+ show_search="filter",
74
+ show_row_numbers=False,
75
+ wrap=True,
76
+ elem_id="sphragis-leaderboard",
77
+ )
78
+ gr.Markdown(METHODOLOGY)
79
+
80
+
81
+ demo.launch()
82
+ </gradio-file>
83
 
84
+ <gradio-file name="results.csv">
85
+ Model,Representation / classifier,Prose 1,Prose 10,Prose 100,Verse sentence 1,Verse sentence 10,Verse sentence 100,Verse metre 1,Verse metre 10,Verse metre 100,Status
86
+ bowphs/GreBerta,"Frozen mean-pooled embeddings + MLP",83.50,96.36,97.69,74.63,80.77,92.51–100.00†,78.85,92.52,96.10‡,"Verse 100 exploratory"
87
+ </gradio-file>
88
+ </gradio-lite>
89
+ </body>
90
+ </html>