felix commited on
Commit
84f536a
·
1 Parent(s): 7411847
Files changed (1) hide show
  1. app.py +33 -4
app.py CHANGED
@@ -4,7 +4,7 @@ import glob
4
  import os
5
  from datetime import datetime
6
 
7
- st.set_page_config(layout="wide")
8
  st.title('Meta Open LLM leaderboard')
9
 
10
  data_path = './data/20230820_0354/'
@@ -16,10 +16,39 @@ parsed_date = datetime.strptime(date_str, "%Y%m%d_%H%M")
16
  # Formatting the parsed date
17
  formatted_date = parsed_date.strftime("%b %d, %Y %H:%M")
18
 
19
- st.subheader("Generated on: " + formatted_date)
 
 
 
 
 
 
 
 
 
 
20
  imgs = glob.glob(os.path.join(data_path, '*.png'))
21
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  cols = st.columns(2)
23
 
24
- for i, img in enumerate(imgs):
25
- cols[i % 2].image(img, width=None)
 
 
 
 
 
 
 
 
4
  import os
5
  from datetime import datetime
6
 
7
+ #st.set_page_config(layout="wide")
8
  st.title('Meta Open LLM leaderboard')
9
 
10
  data_path = './data/20230820_0354/'
 
16
  # Formatting the parsed date
17
  formatted_date = parsed_date.strftime("%b %d, %Y %H:%M")
18
 
19
+ captions_map = {
20
+ "hg_average_to_agentbench_compare.png": "HF to AgentBench compare",
21
+ "hg_average_to_opencompass_compare.png": "HF to OpenCompass compare",
22
+ "hg_average_to_mt_bench_compare.png": "HF to MT-Bench compare",
23
+ "hg_average_to_mosaic_compare.png": "HF to MosaicML compare",
24
+ "hg_average_to_alpacaeval_compare.png": "HF to AlpacaEval compare"
25
+ }
26
+
27
+ st.write("Generated on: " + formatted_date)
28
+ st.divider()
29
+
30
  imgs = glob.glob(os.path.join(data_path, '*.png'))
31
 
32
+ # Extracting images that start with "hf_llm_diagram"
33
+ hf_llm_diagrams = [img for img in imgs if 'hf_llm_diagram' in os.path.basename(img)]
34
+
35
+ # Getting the remaining images
36
+ remaining_imgs = [img for img in imgs if 'hf_llm_diagram' not in os.path.basename(img)]
37
+
38
+ st.write("HuggingFace Open LLM leaderboard by Model Size")
39
+ st.image(hf_llm_diagrams,use_column_width="auto")
40
+
41
+ st.divider()
42
+ st.write("HuggingFace and Other Leaderboards: A Comparative Model Evaluation")
43
+ st.caption("Only models evaluated on both leaderboards are included.")
44
  cols = st.columns(2)
45
 
46
+ for i, img in enumerate(remaining_imgs):
47
+ # Extract the filename from the full image path
48
+ filename = os.path.basename(img)
49
+
50
+ # Get the caption from the captions_map dictionary
51
+ caption = captions_map.get(filename, "") # If no caption is found, it will default to an empty string
52
+
53
+ # Display the image with the caption
54
+ cols[i % 2].image(img, caption=caption, width=None)