Spaces:
Runtime error
Runtime error
Tristan Thrush
commited on
Commit
β’
1060543
1
Parent(s):
3ee9aeb
rolled back latest changes except requirement versions
Browse files
app.py
CHANGED
@@ -130,21 +130,8 @@ dataframe = get_data_wrapper()
|
|
130 |
|
131 |
st.markdown("# π€ Leaderboards")
|
132 |
|
133 |
-
query_params = st.experimental_get_query_params()
|
134 |
-
if "first_query_params" not in st.session_state:
|
135 |
-
st.session_state.first_query_params = query_params
|
136 |
-
first_query_params = st.session_state.first_query_params
|
137 |
-
|
138 |
-
default_task = first_query_params.get("task", [None])[0]
|
139 |
-
default_only_verified = bool(int(first_query_params.get("only_verified", [0])[0]))
|
140 |
-
print(default_only_verified)
|
141 |
-
default_dataset = first_query_params.get("dataset", [None])[0]
|
142 |
-
default_split = first_query_params.get("split", [None])[0]
|
143 |
-
default_config = first_query_params.get("config", [None])[0]
|
144 |
-
|
145 |
only_verified_results = st.sidebar.checkbox(
|
146 |
"Filter for Verified Results",
|
147 |
-
value=default_only_verified,
|
148 |
help="Select this checkbox if you want to see only results produced by the Hugging Face model evaluator, and no self-reported results."
|
149 |
)
|
150 |
|
@@ -152,12 +139,10 @@ selectable_tasks = list(set(dataframe.pipeline_tag))
|
|
152 |
if None in selectable_tasks:
|
153 |
selectable_tasks.remove(None)
|
154 |
selectable_tasks.sort(key=lambda name: name.lower())
|
155 |
-
selectable_tasks = ["-any-"] + selectable_tasks
|
156 |
|
157 |
task = st.sidebar.selectbox(
|
158 |
"Task",
|
159 |
-
selectable_tasks,
|
160 |
-
index=(selectable_tasks).index(default_task) if default_task in selectable_tasks else 0,
|
161 |
help="Filter the selectable datasets by task. Leave as \"-any-\" to see all selectable datasets."
|
162 |
)
|
163 |
|
@@ -168,6 +153,16 @@ selectable_datasets = ["-any-"] + sorted(list(set(dataframe.dataset.tolist())),
|
|
168 |
if "" in selectable_datasets:
|
169 |
selectable_datasets.remove("")
|
170 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
171 |
dataset = st.sidebar.selectbox(
|
172 |
"Dataset",
|
173 |
selectable_datasets,
|
@@ -177,7 +172,7 @@ dataset = st.sidebar.selectbox(
|
|
177 |
|
178 |
dataframe = dataframe[dataframe.only_verified == only_verified_results]
|
179 |
|
180 |
-
st.experimental_set_query_params(**{"dataset": [dataset]
|
181 |
|
182 |
if dataset != "-any-":
|
183 |
dataset_df = dataframe[dataframe.dataset == dataset]
|
@@ -194,7 +189,6 @@ if len(dataset_df) > 0:
|
|
194 |
config = st.sidebar.selectbox(
|
195 |
"Config",
|
196 |
selectable_configs,
|
197 |
-
index=selectable_configs.index(default_config) if default_config in selectable_configs else 0,
|
198 |
help="Filter the results on the current leaderboard by the dataset config. Self-reported results might not report the config, which is why \"-unspecified-\" is an option."
|
199 |
)
|
200 |
dataset_df = dataset_df[dataset_df.config == config]
|
@@ -203,12 +197,8 @@ if len(dataset_df) > 0:
|
|
203 |
split = st.sidebar.selectbox(
|
204 |
"Split",
|
205 |
selectable_splits,
|
206 |
-
index=selectable_splits.index(default_split) if default_split in selectable_splits else 0,
|
207 |
help="Filter the results on the current leaderboard by the dataset split. Self-reported results might not report the split, which is why \"-unspecified-\" is an option."
|
208 |
)
|
209 |
-
|
210 |
-
st.experimental_set_query_params(**{"dataset": [dataset], "only_verified": [int(only_verified_results)], "task": [task], "config": [config], "split": [split]})
|
211 |
-
|
212 |
dataset_df = dataset_df[dataset_df.split == split]
|
213 |
|
214 |
not_selectable_metrics = ["model_id", "dataset", "split", "config", "pipeline_tag", "only_verified"]
|
@@ -292,4 +282,4 @@ if len(dataset_df) > 0:
|
|
292 |
else:
|
293 |
st.markdown(
|
294 |
"No " + ("verified" if only_verified_results else "unverified") + " results to display. Try toggling the verified results filter."
|
295 |
-
)
|
|
|
130 |
|
131 |
st.markdown("# π€ Leaderboards")
|
132 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
133 |
only_verified_results = st.sidebar.checkbox(
|
134 |
"Filter for Verified Results",
|
|
|
135 |
help="Select this checkbox if you want to see only results produced by the Hugging Face model evaluator, and no self-reported results."
|
136 |
)
|
137 |
|
|
|
139 |
if None in selectable_tasks:
|
140 |
selectable_tasks.remove(None)
|
141 |
selectable_tasks.sort(key=lambda name: name.lower())
|
|
|
142 |
|
143 |
task = st.sidebar.selectbox(
|
144 |
"Task",
|
145 |
+
["-any-"] + selectable_tasks,
|
|
|
146 |
help="Filter the selectable datasets by task. Leave as \"-any-\" to see all selectable datasets."
|
147 |
)
|
148 |
|
|
|
153 |
if "" in selectable_datasets:
|
154 |
selectable_datasets.remove("")
|
155 |
|
156 |
+
query_params = st.experimental_get_query_params()
|
157 |
+
if "first_query_params" not in st.session_state:
|
158 |
+
st.session_state.first_query_params = query_params
|
159 |
+
first_query_params = st.session_state.first_query_params
|
160 |
+
|
161 |
+
default_dataset = None
|
162 |
+
if "dataset" in first_query_params:
|
163 |
+
if len(first_query_params["dataset"]) > 0 and first_query_params["dataset"][0] in selectable_datasets:
|
164 |
+
default_dataset = first_query_params["dataset"][0]
|
165 |
+
|
166 |
dataset = st.sidebar.selectbox(
|
167 |
"Dataset",
|
168 |
selectable_datasets,
|
|
|
172 |
|
173 |
dataframe = dataframe[dataframe.only_verified == only_verified_results]
|
174 |
|
175 |
+
st.experimental_set_query_params(**{"dataset": [dataset]})
|
176 |
|
177 |
if dataset != "-any-":
|
178 |
dataset_df = dataframe[dataframe.dataset == dataset]
|
|
|
189 |
config = st.sidebar.selectbox(
|
190 |
"Config",
|
191 |
selectable_configs,
|
|
|
192 |
help="Filter the results on the current leaderboard by the dataset config. Self-reported results might not report the config, which is why \"-unspecified-\" is an option."
|
193 |
)
|
194 |
dataset_df = dataset_df[dataset_df.config == config]
|
|
|
197 |
split = st.sidebar.selectbox(
|
198 |
"Split",
|
199 |
selectable_splits,
|
|
|
200 |
help="Filter the results on the current leaderboard by the dataset split. Self-reported results might not report the split, which is why \"-unspecified-\" is an option."
|
201 |
)
|
|
|
|
|
|
|
202 |
dataset_df = dataset_df[dataset_df.split == split]
|
203 |
|
204 |
not_selectable_metrics = ["model_id", "dataset", "split", "config", "pipeline_tag", "only_verified"]
|
|
|
282 |
else:
|
283 |
st.markdown(
|
284 |
"No " + ("verified" if only_verified_results else "unverified") + " results to display. Try toggling the verified results filter."
|
285 |
+
)
|
utils.py
CHANGED
@@ -13,7 +13,6 @@ metric_ranges = {
|
|
13 |
"accuracy": (0,1),
|
14 |
"precision": (0,1),
|
15 |
"recall": (0,1),
|
16 |
-
"f1": (0,1),
|
17 |
"macro f1": (0,1),
|
18 |
"micro f1": (0,1),
|
19 |
"pearson": (-1, 1),
|
|
|
13 |
"accuracy": (0,1),
|
14 |
"precision": (0,1),
|
15 |
"recall": (0,1),
|
|
|
16 |
"macro f1": (0,1),
|
17 |
"micro f1": (0,1),
|
18 |
"pearson": (-1, 1),
|