Spaces:
Sleeping
Sleeping
Bazsalanszky
commited on
Commit
•
d4be390
1
Parent(s):
2f7f0d5
Test commit
Browse files- src/about.py +11 -33
- src/envs.py +2 -2
src/about.py
CHANGED
@@ -12,8 +12,12 @@ class Task:
|
|
12 |
# ---------------------------------------------------
|
13 |
class Tasks(Enum):
|
14 |
# task_key in the json file, metric_key in the json file, name to display in the leaderboard
|
15 |
-
task0 = Task("
|
16 |
-
task1 = Task("
|
|
|
|
|
|
|
|
|
17 |
|
18 |
NUM_FEWSHOT = 0 # Change with your few shot
|
19 |
# ---------------------------------------------------
|
@@ -21,50 +25,24 @@ NUM_FEWSHOT = 0 # Change with your few shot
|
|
21 |
|
22 |
|
23 |
# Your leaderboard name
|
24 |
-
TITLE = """<h1 align="center" id="space-title">
|
25 |
|
26 |
# What does your leaderboard evaluate?
|
27 |
INTRODUCTION_TEXT = """
|
28 |
-
|
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 |
-
|
37 |
|
38 |
"""
|
39 |
|
40 |
EVALUATION_QUEUE_TEXT = """
|
41 |
-
|
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"
|
|
|
12 |
# ---------------------------------------------------
|
13 |
class Tasks(Enum):
|
14 |
# task_key in the json file, metric_key in the json file, name to display in the leaderboard
|
15 |
+
task0 = Task("category_mean", "history", "History")
|
16 |
+
task1 = Task("category_mean", "grammar", "Grammar")
|
17 |
+
task2 = Task("category_mean", "logic", "Logic")
|
18 |
+
task3 = Task("category_mean", "sayings", "Sayings")
|
19 |
+
task4 = Task("category_mean", "spelling", "Spelling")
|
20 |
+
task5 = Task("category_mean", "Vocabulary", "Vocabulary")
|
21 |
|
22 |
NUM_FEWSHOT = 0 # Change with your few shot
|
23 |
# ---------------------------------------------------
|
|
|
25 |
|
26 |
|
27 |
# Your leaderboard name
|
28 |
+
TITLE = """<h1 align="center" id="space-title">HunBench leaderboard</h1>"""
|
29 |
|
30 |
# What does your leaderboard evaluate?
|
31 |
INTRODUCTION_TEXT = """
|
32 |
+
This leaderboard evaluates the performance of models on the HunBench benchmark. The goal of this benchmark is to evaluate the performance of models on tasks that require a good understanding of the Hungarian language. The benchmark has two key parts. The first one aims to capture the language understanding capabilities of the model, while the second one focuses on the knowledge of the model. The benchmark is divided into several tasks, each evaluating a different aspect of the model's performance. The leaderboard is sorted by the average score of the model on all tasks.
|
33 |
"""
|
34 |
|
35 |
# Which evaluations are you running? how can people reproduce what you have?
|
36 |
LLM_BENCHMARKS_TEXT = f"""
|
37 |
## How it works
|
38 |
+
TODO
|
39 |
## Reproducibility
|
40 |
+
TODO
|
41 |
|
42 |
"""
|
43 |
|
44 |
EVALUATION_QUEUE_TEXT = """
|
45 |
+
TODO
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
"""
|
47 |
|
48 |
CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
|
src/envs.py
CHANGED
@@ -6,12 +6,12 @@ from huggingface_hub import HfApi
|
|
6 |
# ----------------------------------
|
7 |
TOKEN = os.environ.get("HF_TOKEN") # A read/write token for your org
|
8 |
|
9 |
-
OWNER = "
|
10 |
# ----------------------------------
|
11 |
|
12 |
REPO_ID = f"{OWNER}/leaderboard"
|
13 |
QUEUE_REPO = f"{OWNER}/requests"
|
14 |
-
RESULTS_REPO = f"{OWNER}/
|
15 |
|
16 |
# If you setup a cache later, just change HF_HOME
|
17 |
CACHE_PATH=os.getenv("HF_HOME", ".")
|
|
|
6 |
# ----------------------------------
|
7 |
TOKEN = os.environ.get("HF_TOKEN") # A read/write token for your org
|
8 |
|
9 |
+
OWNER = "bazsalanszky" # Change to your org - don't forget to create a results and request dataset, with the correct format!
|
10 |
# ----------------------------------
|
11 |
|
12 |
REPO_ID = f"{OWNER}/leaderboard"
|
13 |
QUEUE_REPO = f"{OWNER}/requests"
|
14 |
+
RESULTS_REPO = f"{OWNER}/hunbench_results"
|
15 |
|
16 |
# If you setup a cache later, just change HF_HOME
|
17 |
CACHE_PATH=os.getenv("HF_HOME", ".")
|