Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
t0-0
commited on
Commit
·
111e1ed
1
Parent(s):
c367864
Remove backend library display from leaderboard and filter
Browse files- app.py +15 -15
- src/display/utils.py +1 -1
- style.css +2 -0
app.py
CHANGED
@@ -78,7 +78,7 @@ def filter_models(
|
|
78 |
add_special_tokens_query: list,
|
79 |
num_few_shots_query: list,
|
80 |
version_query: list,
|
81 |
-
backend_query: list,
|
82 |
) -> pd.DataFrame:
|
83 |
print(f"Initial df shape: {df.shape}")
|
84 |
print(f"Initial df content:\n{df}")
|
@@ -117,8 +117,8 @@ def filter_models(
|
|
117 |
print(f"After version filter: {filtered_df.shape}")
|
118 |
|
119 |
# Backend フィルタリング
|
120 |
-
filtered_df = filtered_df[filtered_df["Backend Library"].isin(backend_query)]
|
121 |
-
print(f"After backend filter: {filtered_df.shape}")
|
122 |
|
123 |
print("Filtered dataframe head:")
|
124 |
print(filtered_df.head())
|
@@ -188,7 +188,7 @@ def update_table(
|
|
188 |
add_special_tokens_query: list,
|
189 |
num_few_shots_query: list,
|
190 |
version_query: list,
|
191 |
-
backend_query: list,
|
192 |
query: str,
|
193 |
):
|
194 |
print(
|
@@ -204,7 +204,7 @@ def update_table(
|
|
204 |
add_special_tokens_query,
|
205 |
num_few_shots_query,
|
206 |
version_query,
|
207 |
-
backend_query,
|
208 |
)
|
209 |
print(f"filtered_df shape after filter_models: {filtered_df.shape}")
|
210 |
|
@@ -251,7 +251,7 @@ leaderboard_df = filter_models(
|
|
251 |
[i.value.name for i in AddSpecialTokens],
|
252 |
[i.value.name for i in NumFewShots],
|
253 |
[i.value.name for i in Version],
|
254 |
-
[i.value.name for i in Backend],
|
255 |
)
|
256 |
|
257 |
leaderboard_df_filtered = filter_models(
|
@@ -262,7 +262,7 @@ leaderboard_df_filtered = filter_models(
|
|
262 |
[i.value.name for i in AddSpecialTokens],
|
263 |
[i.value.name for i in NumFewShots],
|
264 |
[i.value.name for i in Version],
|
265 |
-
[i.value.name for i in Backend],
|
266 |
)
|
267 |
|
268 |
# DataFrameの初期化部分のみを修正
|
@@ -333,12 +333,12 @@ with gr.Blocks() as demo_leaderboard:
|
|
333 |
value=[i.value.name for i in Version],
|
334 |
elem_id="filter-columns-version",
|
335 |
)
|
336 |
-
filter_columns_backend = gr.CheckboxGroup(
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
)
|
342 |
|
343 |
# DataFrameコンポーネントの初期化
|
344 |
leaderboard_table = gr.Dataframe(
|
@@ -371,7 +371,7 @@ with gr.Blocks() as demo_leaderboard:
|
|
371 |
filter_columns_add_special_tokens.change,
|
372 |
filter_columns_num_few_shots.change,
|
373 |
filter_columns_version.change,
|
374 |
-
filter_columns_backend.change,
|
375 |
search_bar.submit,
|
376 |
],
|
377 |
fn=update_table,
|
@@ -384,7 +384,7 @@ with gr.Blocks() as demo_leaderboard:
|
|
384 |
filter_columns_add_special_tokens,
|
385 |
filter_columns_num_few_shots,
|
386 |
filter_columns_version,
|
387 |
-
filter_columns_backend,
|
388 |
search_bar,
|
389 |
],
|
390 |
outputs=leaderboard_table,
|
|
|
78 |
add_special_tokens_query: list,
|
79 |
num_few_shots_query: list,
|
80 |
version_query: list,
|
81 |
+
# backend_query: list,
|
82 |
) -> pd.DataFrame:
|
83 |
print(f"Initial df shape: {df.shape}")
|
84 |
print(f"Initial df content:\n{df}")
|
|
|
117 |
print(f"After version filter: {filtered_df.shape}")
|
118 |
|
119 |
# Backend フィルタリング
|
120 |
+
# filtered_df = filtered_df[filtered_df["Backend Library"].isin(backend_query)]
|
121 |
+
# print(f"After backend filter: {filtered_df.shape}")
|
122 |
|
123 |
print("Filtered dataframe head:")
|
124 |
print(filtered_df.head())
|
|
|
188 |
add_special_tokens_query: list,
|
189 |
num_few_shots_query: list,
|
190 |
version_query: list,
|
191 |
+
# backend_query: list,
|
192 |
query: str,
|
193 |
):
|
194 |
print(
|
|
|
204 |
add_special_tokens_query,
|
205 |
num_few_shots_query,
|
206 |
version_query,
|
207 |
+
# backend_query,
|
208 |
)
|
209 |
print(f"filtered_df shape after filter_models: {filtered_df.shape}")
|
210 |
|
|
|
251 |
[i.value.name for i in AddSpecialTokens],
|
252 |
[i.value.name for i in NumFewShots],
|
253 |
[i.value.name for i in Version],
|
254 |
+
# [i.value.name for i in Backend],
|
255 |
)
|
256 |
|
257 |
leaderboard_df_filtered = filter_models(
|
|
|
262 |
[i.value.name for i in AddSpecialTokens],
|
263 |
[i.value.name for i in NumFewShots],
|
264 |
[i.value.name for i in Version],
|
265 |
+
# [i.value.name for i in Backend],
|
266 |
)
|
267 |
|
268 |
# DataFrameの初期化部分のみを修正
|
|
|
333 |
value=[i.value.name for i in Version],
|
334 |
elem_id="filter-columns-version",
|
335 |
)
|
336 |
+
# filter_columns_backend = gr.CheckboxGroup(
|
337 |
+
# label="Backend Library",
|
338 |
+
# choices=[i.value.name for i in Backend],
|
339 |
+
# value=[i.value.name for i in Backend],
|
340 |
+
# elem_id="filter-columns-backend",
|
341 |
+
# )
|
342 |
|
343 |
# DataFrameコンポーネントの初期化
|
344 |
leaderboard_table = gr.Dataframe(
|
|
|
371 |
filter_columns_add_special_tokens.change,
|
372 |
filter_columns_num_few_shots.change,
|
373 |
filter_columns_version.change,
|
374 |
+
# filter_columns_backend.change,
|
375 |
search_bar.submit,
|
376 |
],
|
377 |
fn=update_table,
|
|
|
384 |
filter_columns_add_special_tokens,
|
385 |
filter_columns_num_few_shots,
|
386 |
filter_columns_version,
|
387 |
+
# filter_columns_backend,
|
388 |
search_bar,
|
389 |
],
|
390 |
outputs=leaderboard_table,
|
src/display/utils.py
CHANGED
@@ -56,7 +56,7 @@ auto_eval_column_dict.append(["add_special_tokens", ColumnContent, ColumnContent
|
|
56 |
auto_eval_column_dict.append(
|
57 |
["llm_jp_eval_version", ColumnContent, ColumnContent("llm-jp-eval version", "str", False)]
|
58 |
)
|
59 |
-
auto_eval_column_dict.append(["backend", ColumnContent, ColumnContent("Backend Library", "str", False)])
|
60 |
auto_eval_column_dict.append(["dummy", ColumnContent, ColumnContent("model_name_for_query", "str", False, dummy=True)])
|
61 |
|
62 |
# We use make dataclass to dynamically fill the scores from Tasks
|
|
|
56 |
auto_eval_column_dict.append(
|
57 |
["llm_jp_eval_version", ColumnContent, ColumnContent("llm-jp-eval version", "str", False)]
|
58 |
)
|
59 |
+
auto_eval_column_dict.append(["backend", ColumnContent, ColumnContent("Backend Library", "str", False, dummy=True)])
|
60 |
auto_eval_column_dict.append(["dummy", ColumnContent, ColumnContent("model_name_for_query", "str", False, dummy=True)])
|
61 |
|
62 |
# We use make dataclass to dynamically fill the scores from Tasks
|
style.css
CHANGED
@@ -6,6 +6,8 @@
|
|
6 |
}
|
7 |
|
8 |
/* Hides the final AutoEvalColumn */
|
|
|
|
|
9 |
#llm-benchmark-tab-table table td:last-child,
|
10 |
#llm-benchmark-tab-table table th:last-child {
|
11 |
display: none;
|
|
|
6 |
}
|
7 |
|
8 |
/* Hides the final AutoEvalColumn */
|
9 |
+
#llm-benchmark-tab-table table td:nth-last-child(2),
|
10 |
+
#llm-benchmark-tab-table table th:nth-last-child(2),
|
11 |
#llm-benchmark-tab-table table td:last-child,
|
12 |
#llm-benchmark-tab-table table th:last-child {
|
13 |
display: none;
|