huckiyang commited on
Commit
e9a9bc8
·
1 Parent(s): a25014e

[data] fixing visual

Browse files
Files changed (1) hide show
  1. src/about.py +28 -33
src/about.py CHANGED
@@ -21,52 +21,47 @@ NUM_FEWSHOT = 0 # Change with your few shot
21
 
22
 
23
  # Your leaderboard name
24
- TITLE = """<h1 align="center" id="space-title">🪷 LOTUS: A Leaderboard for Detailed Image Captioning from Quality to Societal Bias and User Preferences</h1>"""
 
 
 
 
25
 
26
  # What does your leaderboard evaluate?
27
  INTRODUCTION_TEXT = """
28
- (ACL 2025 Industry Track) Large Vision-Language Models (LVLMs) have transformed image captioning, shifting from concise captions to detailed, context-rich descriptions. We introduce LOTUS, a unified leaderboard for evaluating such detailed captions, addressing three main gaps in existing evaluation approaches: lack of standardized criteria, absence of bias-aware assessments, and evaluations that disregard user preferences.
29
- """
30
 
31
- # Which evaluations are you running? how can people reproduce what you have?
32
- LLM_BENCHMARKS_TEXT = f"""
33
- ## How it works
34
 
35
- ## Reproducibility
36
- To reproduce our results, here is the commands you can run:
37
 
38
  """
39
 
40
- EVALUATION_QUEUE_TEXT = """
41
- ## Some good practices before submitting a model
42
-
43
- ### 1) Make sure you can load your model and tokenizer using AutoClasses:
44
- ```python
45
- from transformers import AutoConfig, AutoModel, AutoTokenizer
46
- config = AutoConfig.from_pretrained("your model name", revision=revision)
47
- model = AutoModel.from_pretrained("your model name", revision=revision)
48
- tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)
49
- ```
50
- If this step fails, follow the error messages to debug your model before submitting it. It's likely your model has been improperly uploaded.
51
 
52
- Note: make sure your model is public!
53
- Note: if your model needs `use_remote_code=True`, we do not support this option yet but we are working on adding it, stay posted!
54
 
55
- ### 2) Convert your model weights to [safetensors](https://huggingface.co/docs/safetensors/index)
56
- It's a new format for storing weights which is safer and faster to load and use. It will also allow us to add the number of parameters of your model to the `Extended Viewer`!
57
 
58
- ### 3) Make sure your model has an open license!
59
- This is a leaderboard for Open LLMs, and we'd love for as many people as possible to know they can use your model 🤗
60
 
61
- ### 4) Fill up your model card
62
- When we add extra information about models to the leaderboard, it will be automatically taken from the model card
 
63
 
64
- ## In case of model failure
65
- If your model is displayed in the `FAILED` category, its execution stopped.
66
- Make sure you have followed the above steps first.
67
- If everything is done, check you can launch the EleutherAIHarness on your model locally, using the above command without modifications (you can add `--limit` to limit the number of examples per task).
68
  """
69
 
70
- CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
71
- CITATION_BUTTON_TEXT = r"""
 
 
 
 
 
 
 
72
  """
 
21
 
22
 
23
  # Your leaderboard name
24
+ TITLE = """
25
+ <div align="center">
26
+ <h1>🪷 LOTUS: Detailed LVLM Evaluation from Quality to Societal Bias</h1>
27
+ </div>
28
+ """
29
 
30
  # What does your leaderboard evaluate?
31
  INTRODUCTION_TEXT = """
 
 
32
 
33
+ **LOTUS** (short for **L**arge Language Model (**L**L**M**) E**o**s E**t**hos and Pa**u**to**s** Benchmark) is a new comprehensive benchmark for Large Vision Language Models (LVLMs).
 
 
34
 
35
+ For more information, check out our [paper](https://arxiv.org/abs/2402.10542) and the [project page](https://lotus-benchmark.github.io/).
 
36
 
37
  """
38
 
39
+ # Which evaluations are you running? how can people reproduce what you have?
40
+ LLM_BENCHMARKS_TEXT = """
41
+ Details about the LLM benchmarks will go here.
42
+ """
 
 
 
 
 
 
 
43
 
44
+ EVALUATION_QUEUE_TEXT = """
45
+ ## Evaluation Queue and Submission
46
 
47
+ Models are evaluated on the private test set of COCO Captions, and results are protected from misuse.
 
48
 
49
+ **To submit your model for evaluation, please follow these steps:**
 
50
 
51
+ 1. **Prepare your model's predictions**: Ensure your model's predictions are in the same format as the COCO Captions [dataset](https://cocodataset.org/#format-data).
52
+ 2. **Submit via our form**: Fill out the [submission form](https://forms.gle/your_form_link_here) with your model details and prediction file.
53
+ 3. **Wait for results**: The evaluation may take some time. Results will be updated on the leaderboard.
54
 
55
+ For any issues or questions, please open an issue on our [GitHub repository](https://github.com/lotus-benchmark/lotus).
 
 
 
56
  """
57
 
58
+ CITATION_BUTTON_LABEL = "BibTeX" # Keep this label or change if needed
59
+
60
+ CITATION_BUTTON_TEXT = """
61
+ @inproceedings{hirota2025lotus,
62
+ author = {Yusuke Hirota and Boyi Li and Ryo Hachiuma and Yueh-Hua Wu and Boris Ivanovic and Yuta Nakashima and Marco Pavone and Yejin Choi and Yu-Chiang Frank Wang and Huck Yang},
63
+ title = {LOTUS: {A} Leaderboard for Detailed Image Captioning from Quality to Societal Bias and User Preferences},
64
+ booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL)},
65
+ year = {2025}
66
+ }
67
  """