Spaces:
Runtime error
Runtime error
Add description
Browse files
README.md
CHANGED
|
@@ -1,9 +1,10 @@
|
|
| 1 |
---
|
| 2 |
-
title: Gem
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: pink
|
| 5 |
colorTo: indigo
|
| 6 |
sdk: streamlit
|
|
|
|
| 7 |
app_file: app.py
|
| 8 |
pinned: false
|
| 9 |
---
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Gem Submissions
|
| 3 |
+
emoji: π
|
| 4 |
colorFrom: pink
|
| 5 |
colorTo: indigo
|
| 6 |
sdk: streamlit
|
| 7 |
+
sdk_version: 1.0.0
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
---
|
app.py
CHANGED
|
@@ -45,7 +45,24 @@ def http_post(
|
|
| 45 |
return response
|
| 46 |
|
| 47 |
|
| 48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
with st.form(key="form"):
|
| 50 |
# Flush local repo
|
| 51 |
shutil.rmtree(LOCAL_REPO, ignore_errors=True)
|
|
@@ -78,8 +95,10 @@ with st.form(key="form"):
|
|
| 78 |
In this case, `dataset_identifier` is the identifier of the dataset
|
| 79 |
followed by an identifier of the set the outputs were created from, for
|
| 80 |
example `_validation` or `_test`. For example, the `mlsum_de` test set
|
| 81 |
-
would have the identifier `mlsum_de_test`.
|
| 82 |
-
|
|
|
|
|
|
|
| 83 |
"""
|
| 84 |
)
|
| 85 |
with open("sample-submission.json", "r") as f:
|
|
@@ -87,9 +106,9 @@ with st.form(key="form"):
|
|
| 87 |
st.json(example_submission)
|
| 88 |
|
| 89 |
token = st.text_input(
|
| 90 |
-
"
|
| 91 |
type="password",
|
| 92 |
-
help="You can generate an access token via your Hub settings. See the [docs](https://huggingface.co/docs/hub/security#user-access-tokens) for more details",
|
| 93 |
)
|
| 94 |
|
| 95 |
submit_button = st.form_submit_button("Make Submission")
|
|
|
|
| 45 |
return response
|
| 46 |
|
| 47 |
|
| 48 |
+
###########
|
| 49 |
+
### APP ###
|
| 50 |
+
###########
|
| 51 |
+
st.title("GEM Submissions")
|
| 52 |
+
st.markdown(
|
| 53 |
+
"""
|
| 54 |
+
Welcome to the [GEM benchmark](https://gem-benchmark.com/)! GEM is a benchmark environment for Natural Language Generation with a focus on its Evaluation, both through human annotations and automated Metrics.
|
| 55 |
+
|
| 56 |
+
GEM aims to:
|
| 57 |
+
|
| 58 |
+
- measure NLG progress across many NLG tasks across languages.
|
| 59 |
+
- audit data and models and present results via data cards and model robustness reports.
|
| 60 |
+
- develop standards for evaluation of generated text using both automated and human metrics.
|
| 61 |
+
|
| 62 |
+
Use this page to submit your predcitions to the benchmark.
|
| 63 |
+
"""
|
| 64 |
+
)
|
| 65 |
+
|
| 66 |
with st.form(key="form"):
|
| 67 |
# Flush local repo
|
| 68 |
shutil.rmtree(LOCAL_REPO, ignore_errors=True)
|
|
|
|
| 95 |
In this case, `dataset_identifier` is the identifier of the dataset
|
| 96 |
followed by an identifier of the set the outputs were created from, for
|
| 97 |
example `_validation` or `_test`. For example, the `mlsum_de` test set
|
| 98 |
+
would have the identifier `mlsum_de_test`. The `keys` field can be set
|
| 99 |
+
to avoid accidental shuffling to impact your metrics. Simply add a list
|
| 100 |
+
of the `gem_id` for each output example in the same order as your
|
| 101 |
+
values. Please see the sample submission below:
|
| 102 |
"""
|
| 103 |
)
|
| 104 |
with open("sample-submission.json", "r") as f:
|
|
|
|
| 106 |
st.json(example_submission)
|
| 107 |
|
| 108 |
token = st.text_input(
|
| 109 |
+
"Enter π€ Hub access token",
|
| 110 |
type="password",
|
| 111 |
+
help="You can generate an access token via your π€ Hub settings. See the [docs](https://huggingface.co/docs/hub/security#user-access-tokens) for more details",
|
| 112 |
)
|
| 113 |
|
| 114 |
submit_button = st.form_submit_button("Make Submission")
|