Update app.py
Browse files
app.py
CHANGED
@@ -3,58 +3,65 @@ from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
|
|
3 |
import pandas as pd
|
4 |
from apscheduler.schedulers.background import BackgroundScheduler
|
5 |
from huggingface_hub import snapshot_download
|
6 |
-
|
7 |
from src.about import (
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
)
|
15 |
from src.display.css_html_js import custom_css
|
16 |
from src.display.utils import (
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
)
|
27 |
from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN
|
28 |
from src.populate import get_evaluation_queue_df, get_leaderboard_df
|
29 |
from src.submission.submit import add_new_eval
|
30 |
|
31 |
-
|
32 |
def restart_space():
|
33 |
-
|
34 |
|
35 |
### Space initialisation
|
36 |
try:
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
|
|
|
|
|
|
|
|
|
|
41 |
except Exception:
|
42 |
-
|
|
|
43 |
try:
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
|
|
|
|
|
|
|
|
|
|
48 |
except Exception:
|
49 |
-
|
50 |
-
|
51 |
|
52 |
LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS)
|
53 |
-
|
54 |
(
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
|
59 |
|
60 |
def init_leaderboard(dataframe):
|
@@ -70,117 +77,40 @@ def init_leaderboard(dataframe):
|
|
70 |
),
|
71 |
search_columns=[AutoEvalColumn.model.name, AutoEvalColumn.license.name],
|
72 |
hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden],
|
73 |
-
filter_columns=[
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
|
|
|
|
83 |
user_data = data.load_data()
|
84 |
if user_id in user_data:
|
85 |
-
|
86 |
else:
|
87 |
-
|
88 |
-
import gradio as gr # No indentation before this line!
|
89 |
-
from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
|
90 |
-
import pandas as pd
|
91 |
-
# ... (Rest of your code)
|
92 |
-
revision_name_textbox = gr.Textbox(label="Revision commit", placeholder="main")
|
93 |
-
model_type = gr.Dropdown(
|
94 |
-
choices=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown],
|
95 |
-
label="Model type",
|
96 |
-
multiselect=False,
|
97 |
-
value=None,
|
98 |
-
interactive=True,
|
99 |
-
)
|
100 |
-
with gr.Column():
|
101 |
-
precision = gr.Dropdown(
|
102 |
-
choices=[i.value.name for i in Precision if i != Precision.Unknown],
|
103 |
-
label="Precision",
|
104 |
-
multiselect=False,
|
105 |
-
value="float16",
|
106 |
-
interactive=True,
|
107 |
-
)
|
108 |
-
weight_type = gr.Dropdown(
|
109 |
-
choices=[i.value.name for i in WeightType],
|
110 |
-
label="Weights type",
|
111 |
-
multiselect=False,
|
112 |
-
value="Original",
|
113 |
-
interactive=True,
|
114 |
-
)
|
115 |
-
base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)")
|
116 |
-
model_name_textbox = gr.Textbox(label="Model name") # Add this line
|
117 |
-
|
118 |
-
submit_button = gr.Button("Submit Eval")
|
119 |
-
submission_result = gr.Markdown()
|
120 |
|
121 |
-
# Indent the entire `with gr.Row()` section
|
122 |
-
with gr.Row():
|
123 |
-
with gr.Accordion("π Citation", open=False):
|
124 |
-
citation_button = gr.Textbox(
|
125 |
-
value=CITATION_BUTTON_TEXT,
|
126 |
-
label=CITATION_BUTTON_LABEL,
|
127 |
-
lines=20,
|
128 |
-
elem_id="citation-button",
|
129 |
-
show_copy_button=True,
|
130 |
-
)
|
131 |
-
|
132 |
-
# ... (Rest of your code)
|
133 |
-
|
134 |
-
submit_button.click(
|
135 |
-
add_new_eval,
|
136 |
-
[
|
137 |
-
model_name_textbox,
|
138 |
-
base_model_name_textbox,
|
139 |
-
revision_name_textbox,
|
140 |
-
precision,
|
141 |
-
weight_type,
|
142 |
-
model_type,
|
143 |
-
],
|
144 |
-
submission_result,
|
145 |
-
)
|
146 |
-
|
147 |
-
# ... (Scheduler)
|
148 |
-
start_button = gr.Button("Start", elem_id="start_button")
|
149 |
-
claim_image = gr.Image(value="path/to/your/claim_image.png", interactive=True, elem_id="claim_image")
|
150 |
-
start_button.click(fn=display_user_data, inputs=[gr.inputs.Textbox(label="Enter your Telegram User ID")], outputs=user_data_output)
|
151 |
-
claim_image.click(fn=lambda: "You have claimed your reward!", outputs=gr.outputs.Textbox(label="Message"))
|
152 |
-
iface = gr.Interface(
|
153 |
-
fn=None, # You can add functions for tasks later
|
154 |
-
inputs=[
|
155 |
-
join_channel_section,
|
156 |
-
invite_friends_section,
|
157 |
-
],
|
158 |
-
outputs=[user_data_output],
|
159 |
-
title="Mukera",
|
160 |
-
)
|
161 |
-
iface.launch(share=True, inline=True)
|
162 |
demo = gr.Blocks(css=custom_css)
|
|
|
163 |
with demo:
|
164 |
gr.HTML(TITLE)
|
165 |
gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text")
|
166 |
-
|
167 |
with gr.Tabs(elem_classes="tab-buttons") as tabs:
|
168 |
with gr.TabItem("π
LLM Benchmark", elem_id="llm-benchmark-tab-table", id=0):
|
169 |
leaderboard = init_leaderboard(LEADERBOARD_DF)
|
170 |
-
|
171 |
with gr.TabItem("π About", elem_id="llm-benchmark-tab-table", id=2):
|
172 |
gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
|
173 |
-
|
174 |
with gr.TabItem("π Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
|
175 |
with gr.Column():
|
176 |
with gr.Row():
|
177 |
gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text")
|
178 |
-
|
179 |
with gr.Column():
|
180 |
-
with gr.Accordion(
|
181 |
-
f"β
Finished Evaluations ({len(finished_eval_queue_df)})",
|
182 |
-
open=False,
|
183 |
-
):
|
184 |
with gr.Row():
|
185 |
finished_eval_table = gr.components.Dataframe(
|
186 |
value=finished_eval_queue_df,
|
@@ -188,22 +118,14 @@ with demo:
|
|
188 |
datatype=EVAL_TYPES,
|
189 |
row_count=5,
|
190 |
)
|
191 |
-
with gr.Accordion(
|
192 |
-
f"π Running Evaluation Queue ({len(running_eval_queue_df)})",
|
193 |
-
open=False,
|
194 |
-
):
|
195 |
with gr.Row():
|
196 |
-
running_eval_table = gr.components.Dataframe(
|
197 |
-
value=running_eval_queue_df,
|
198 |
headers=EVAL_COLS,
|
199 |
datatype=EVAL_TYPES,
|
200 |
row_count=5,
|
201 |
)
|
202 |
-
|
203 |
-
with gr.Accordion(
|
204 |
-
f"β³ Pending Evaluation Queue ({len(pending_eval_queue_df)})",
|
205 |
-
open=False,
|
206 |
-
):
|
207 |
with gr.Row():
|
208 |
pending_eval_table = gr.components.Dataframe(
|
209 |
value=pending_eval_queue_df,
|
@@ -211,40 +133,47 @@ with demo:
|
|
211 |
datatype=EVAL_TYPES,
|
212 |
row_count=5,
|
213 |
)
|
214 |
-
|
215 |
-
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
223 |
-
|
224 |
-
|
225 |
-
|
226 |
-
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
233 |
-
|
234 |
-
|
235 |
-
|
236 |
-
|
237 |
-
|
238 |
-
|
239 |
-
|
240 |
-
|
241 |
-
|
242 |
-
|
243 |
-
|
244 |
-
base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)")
|
245 |
-
|
246 |
submit_button = gr.Button("Submit Eval")
|
247 |
submission_result = gr.Markdown()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
248 |
submit_button.click(
|
249 |
add_new_eval,
|
250 |
[
|
@@ -257,18 +186,8 @@ with demo:
|
|
257 |
],
|
258 |
submission_result,
|
259 |
)
|
260 |
-
|
261 |
-
|
262 |
-
|
263 |
-
|
264 |
-
|
265 |
-
label=CITATION_BUTTON_LABEL,
|
266 |
-
lines=20,
|
267 |
-
elem_id="citation-button",
|
268 |
-
show_copy_button=True,
|
269 |
-
)
|
270 |
-
|
271 |
-
scheduler = BackgroundScheduler()
|
272 |
-
scheduler.add_job(restart_space, "interval", seconds=1800)
|
273 |
-
scheduler.start()
|
274 |
-
demo.queue(default_concurrency_limit=40).launch()
|
|
|
3 |
import pandas as pd
|
4 |
from apscheduler.schedulers.background import BackgroundScheduler
|
5 |
from huggingface_hub import snapshot_download
|
|
|
6 |
from src.about import (
|
7 |
+
CITATION_BUTTON_LABEL,
|
8 |
+
CITATION_BUTTON_TEXT,
|
9 |
+
EVALUATION_QUEUE_TEXT,
|
10 |
+
INTRODUCTION_TEXT,
|
11 |
+
LLM_BENCHMARKS_TEXT,
|
12 |
+
TITLE,
|
13 |
)
|
14 |
from src.display.css_html_js import custom_css
|
15 |
from src.display.utils import (
|
16 |
+
BENCHMARK_COLS,
|
17 |
+
COLS,
|
18 |
+
EVAL_COLS,
|
19 |
+
EVAL_TYPES,
|
20 |
+
AutoEvalColumn,
|
21 |
+
ModelType,
|
22 |
+
fields,
|
23 |
+
WeightType,
|
24 |
+
Precision,
|
25 |
)
|
26 |
from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN
|
27 |
from src.populate import get_evaluation_queue_df, get_leaderboard_df
|
28 |
from src.submission.submit import add_new_eval
|
29 |
|
|
|
30 |
def restart_space():
|
31 |
+
API.restart_space(repo_id=REPO_ID)
|
32 |
|
33 |
### Space initialisation
|
34 |
try:
|
35 |
+
print(EVAL_REQUESTS_PATH)
|
36 |
+
snapshot_download(
|
37 |
+
repo_id=QUEUE_REPO,
|
38 |
+
local_dir=EVAL_REQUESTS_PATH,
|
39 |
+
repo_type="dataset",
|
40 |
+
tqdm_class=None,
|
41 |
+
etag_timeout=30,
|
42 |
+
token=TOKEN,
|
43 |
+
)
|
44 |
except Exception:
|
45 |
+
restart_space()
|
46 |
+
|
47 |
try:
|
48 |
+
print(EVAL_RESULTS_PATH)
|
49 |
+
snapshot_download(
|
50 |
+
repo_id=RESULTS_REPO,
|
51 |
+
local_dir=EVAL_RESULTS_PATH,
|
52 |
+
repo_type="dataset",
|
53 |
+
tqdm_class=None,
|
54 |
+
etag_timeout=30,
|
55 |
+
token=TOKEN,
|
56 |
+
)
|
57 |
except Exception:
|
58 |
+
restart_space()
|
|
|
59 |
|
60 |
LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS)
|
|
|
61 |
(
|
62 |
+
finished_eval_queue_df,
|
63 |
+
running_eval_queue_df,
|
64 |
+
pending_eval_queue_df,
|
65 |
) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
|
66 |
|
67 |
def init_leaderboard(dataframe):
|
|
|
77 |
),
|
78 |
search_columns=[AutoEvalColumn.model.name, AutoEvalColumn.license.name],
|
79 |
hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden],
|
80 |
+
filter_columns=[ ColumnFilter(AutoEvalColumn.model_type.name, type="checkboxgroup", label="Model types"),
|
81 |
+
ColumnFilter(AutoEvalColumn.precision.name, type="checkboxgroup", label="Precision"),
|
82 |
+
ColumnFilter(
|
83 |
+
AutoEvalColumn.params.name, type="slider", min=0.01, max=150, label="Select the number of parameters (B)"
|
84 |
+
),
|
85 |
+
ColumnFilter(AutoEvalColumn.still_on_hub.name, type="boolean", label="Deleted/incomplete", default=True),
|
86 |
+
],
|
87 |
+
bool_checkboxgroup_label="Hide models",
|
88 |
+
interactive=False,
|
89 |
+
)
|
90 |
+
|
91 |
+
def display_user_data(user_id):
|
92 |
user_data = data.load_data()
|
93 |
if user_id in user_data:
|
94 |
+
return f"Points: {user_data[user_id]['points']}\nReferrals: {len(user_data[user_id]['referrals'])}"
|
95 |
else:
|
96 |
+
return "User not found"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
97 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
98 |
demo = gr.Blocks(css=custom_css)
|
99 |
+
|
100 |
with demo:
|
101 |
gr.HTML(TITLE)
|
102 |
gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text")
|
|
|
103 |
with gr.Tabs(elem_classes="tab-buttons") as tabs:
|
104 |
with gr.TabItem("π
LLM Benchmark", elem_id="llm-benchmark-tab-table", id=0):
|
105 |
leaderboard = init_leaderboard(LEADERBOARD_DF)
|
|
|
106 |
with gr.TabItem("π About", elem_id="llm-benchmark-tab-table", id=2):
|
107 |
gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
|
|
|
108 |
with gr.TabItem("π Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
|
109 |
with gr.Column():
|
110 |
with gr.Row():
|
111 |
gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text")
|
|
|
112 |
with gr.Column():
|
113 |
+
with gr.Accordion(f"β
Finished Evaluations ({len(finished_eval_queue_df)})", open=False):
|
|
|
|
|
|
|
114 |
with gr.Row():
|
115 |
finished_eval_table = gr.components.Dataframe(
|
116 |
value=finished_eval_queue_df,
|
|
|
118 |
datatype=EVAL_TYPES,
|
119 |
row_count=5,
|
120 |
)
|
121 |
+
with gr.Accordion(f"π Running Evaluation Queue ({len(running_eval_queue_df)})", open=False):
|
|
|
|
|
|
|
122 |
with gr.Row():
|
123 |
+
running_eval_table = gr.components.Dataframe( value=running_eval_queue_df,
|
|
|
124 |
headers=EVAL_COLS,
|
125 |
datatype=EVAL_TYPES,
|
126 |
row_count=5,
|
127 |
)
|
128 |
+
with gr.Accordion(f"β³ Pending Evaluation Queue ({len(pending_eval_queue_df)})", open=False):
|
|
|
|
|
|
|
|
|
129 |
with gr.Row():
|
130 |
pending_eval_table = gr.components.Dataframe(
|
131 |
value=pending_eval_queue_df,
|
|
|
133 |
datatype=EVAL_TYPES,
|
134 |
row_count=5,
|
135 |
)
|
136 |
+
with gr.Row():
|
137 |
+
gr.Markdown("# βοΈβ¨ Submit your model here!", elem_classes="markdown-text")
|
138 |
+
with gr.Row():
|
139 |
+
with gr.Column():
|
140 |
+
revision_name_textbox = gr.Textbox(label="Revision commit", placeholder="main")
|
141 |
+
model_type = gr.Dropdown(
|
142 |
+
choices=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown],
|
143 |
+
label="Model type",
|
144 |
+
multiselect=False,
|
145 |
+
value=None,
|
146 |
+
interactive=True,
|
147 |
+
)
|
148 |
+
with gr.Column():
|
149 |
+
precision = gr.Dropdown(
|
150 |
+
choices=[i.value.name for i in Precisi
|
151 |
+
on if i != Precision.Unknown],
|
152 |
+
label="Precision",
|
153 |
+
multiselect=False,
|
154 |
+
value="float16",
|
155 |
+
interactive=True,
|
156 |
+
)
|
157 |
+
weight_type = gr.Dropdown(
|
158 |
+
choices=[i.value.name for i in WeightType],
|
159 |
+
label="Weights type",
|
160 |
+
multiselect=False,
|
161 |
+
value="Original",
|
162 |
+
interactive=True,
|
163 |
+
)
|
164 |
+
base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)")
|
165 |
+
model_name_textbox = gr.Textbox(label="Model name") # Add this line
|
|
|
|
|
166 |
submit_button = gr.Button("Submit Eval")
|
167 |
submission_result = gr.Markdown()
|
168 |
+
with gr.Row():
|
169 |
+
with gr.Accordion("π Citation", open=False):
|
170 |
+
citation_button = gr.Textbox(
|
171 |
+
value=CITATION_BUTTON_TEXT,
|
172 |
+
label=CITATION_BUTTON_LABEL,
|
173 |
+
lines=20,
|
174 |
+
elem_id="citation-button",
|
175 |
+
show_copy_button=True,
|
176 |
+
)
|
177 |
submit_button.click(
|
178 |
add_new_eval,
|
179 |
[
|
|
|
186 |
],
|
187 |
submission_result,
|
188 |
)
|
189 |
+
start_button = gr.Button("Start", elem_id="start_button")
|
190 |
+
scheduler = BackgroundScheduler()
|
191 |
+
scheduler.add_job(restart_space, "interval", seconds=1800)
|
192 |
+
scheduler.start()
|
193 |
+
demo.queue(default_concurrency_limit=40).launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|