Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,83 +1,11 @@
|
|
1 |
import pandas as pd
|
2 |
import gradio as gr
|
3 |
import os
|
4 |
-
from gradio_rangeslider import RangeSlider
|
5 |
|
6 |
-
|
7 |
-
|
8 |
-
# Main Leaderboard containing everything
|
9 |
-
text_leaderboard = pd.read_csv(os.path.join('assets', 'merged_data.csv'))
|
10 |
-
text_leaderboard = text_leaderboard.sort_values(by='Clemscore', ascending=False)
|
11 |
-
|
12 |
-
open_weight_df = text_leaderboard[text_leaderboard['Open Weight'] == True]
|
13 |
-
if not open_weight_df.empty: # Check if filtered df is non-empty
|
14 |
-
max_parameter_size = open_weight_df['Parameters (B)'].max()
|
15 |
-
|
16 |
-
# Short leaderboard containing fixed columns
|
17 |
-
short_leaderboard = filter_cols(text_leaderboard)
|
18 |
-
|
19 |
-
## Extract data
|
20 |
-
langs = []
|
21 |
-
licenses = []
|
22 |
-
ip_prices = []
|
23 |
-
op_prices = []
|
24 |
-
latencies = []
|
25 |
-
parameters = []
|
26 |
-
contexts = []
|
27 |
-
dates = []
|
28 |
-
|
29 |
-
for i in range(len(text_leaderboard)):
|
30 |
-
lang_splits = text_leaderboard.iloc[i]['Languages'].split(',')
|
31 |
-
lang_splits = [s.strip() for s in lang_splits]
|
32 |
-
langs += lang_splits
|
33 |
-
license_name = text_leaderboard.iloc[i]['License Name']
|
34 |
-
|
35 |
-
licenses.append(license_name)
|
36 |
-
ip_prices.append(text_leaderboard.iloc[i]['Input $/1M tokens'])
|
37 |
-
op_prices.append(text_leaderboard.iloc[i]['Output $/1M tokens'])
|
38 |
-
latencies.append(text_leaderboard.iloc[i]['Latency (s)'])
|
39 |
-
parameters.append(text_leaderboard.iloc[i]['Parameters (B)'])
|
40 |
-
contexts.append(text_leaderboard.iloc[i]['Context Size (k)'])
|
41 |
-
dates.append(text_leaderboard.iloc[i]['Release Date'])
|
42 |
-
|
43 |
-
|
44 |
-
langs = list(set(langs))
|
45 |
-
langs.sort()
|
46 |
-
|
47 |
-
licenses = list(set(licenses))
|
48 |
-
licenses.sort()
|
49 |
-
|
50 |
-
max_input_price = max(ip_prices)
|
51 |
-
max_output_price = max(op_prices)
|
52 |
-
max_latency = max(latencies)
|
53 |
-
|
54 |
-
min_parameters = 0 if pd.isna(min(parameters)) else min(parameters)
|
55 |
-
max_parameter = max_parameter_size
|
56 |
-
parameter_step = 1
|
57 |
-
print(f"MIN {min_parameters}, MAX {max_parameter}")
|
58 |
-
|
59 |
-
min_context = min(contexts)
|
60 |
-
max_context = max(contexts)
|
61 |
-
context_step = 8
|
62 |
-
|
63 |
-
min_date = min(dates)
|
64 |
-
max_date = max(dates)
|
65 |
-
|
66 |
-
TITLE = """<h1 align="center" id="space-title"> LLM Calculator ⚖️⚡ 📏💰</h1>"""
|
67 |
-
CSS = """
|
68 |
-
#double-slider-1 {height: 100px}
|
69 |
-
#double-slider-2 {height: 100px}
|
70 |
-
#double-slider-3 {height: 100px}
|
71 |
-
#double-slider-4 {height: 100px}
|
72 |
-
"""
|
73 |
-
|
74 |
-
llm_calc_app = gr.Blocks(css=CSS)
|
75 |
with llm_calc_app:
|
76 |
|
77 |
-
gr.HTML(TITLE)
|
78 |
-
|
79 |
##################################################
|
80 |
-
|
81 |
with gr.Row():
|
82 |
|
83 |
#####################################
|
@@ -86,14 +14,6 @@ with llm_calc_app:
|
|
86 |
## Language Select
|
87 |
with gr.Column():
|
88 |
|
89 |
-
with gr.Row():
|
90 |
-
lang_dropdown = gr.Dropdown(
|
91 |
-
choices=langs,
|
92 |
-
value=[],
|
93 |
-
multiselect=True,
|
94 |
-
label="Select Languages 🗣️"
|
95 |
-
)
|
96 |
-
|
97 |
with gr.Row():
|
98 |
start_date = gr.DateTime(
|
99 |
)
|
@@ -106,37 +26,3 @@ with llm_calc_app:
|
|
106 |
llm_calc_app.queue()
|
107 |
llm_calc_app.launch()
|
108 |
|
109 |
-
|
110 |
-
|
111 |
-
"""
|
112 |
-
model_name, input_price, output_price,
|
113 |
-
multimodality_image,multimodality_multiple_image,multimodality_audio,multimodality_video,
|
114 |
-
source,licence_name,licence_url,languages,release_date,
|
115 |
-
parameters_estimated,parameters_actual,
|
116 |
-
|
117 |
-
open_weight,context,
|
118 |
-
|
119 |
-
additional_prices_context_caching,
|
120 |
-
additional_prices_context_storage,
|
121 |
-
additional_prices_image_input,additional_prices_image_output,additional_prices_video_input,additional_prices_video_output,additional_prices_audio_input,additional_prices_audio_output,clemscore_v1.6.5_multimodal,clemscore_v1.6.5_ascii,clemscore_v1.6,latency_v1.6,latency_v1.6.5_multimodal,latency_v1.6.5_ascii,
|
122 |
-
|
123 |
-
average_clemscore,average_latency,parameters
|
124 |
-
|
125 |
-
Final list
|
126 |
-
|
127 |
-
model_name, input_price, output_price,
|
128 |
-
multimodality_image,multimodality_multiple_image,multimodality_audio,multimodality_video,
|
129 |
-
source,licence_name,licence_url,languages,release_date, open_weight,context, average_clemscore,average_latency,parameters
|
130 |
-
|
131 |
-
|
132 |
-
Filter
|
133 |
-
multimodality_image,multimodality_multiple_image,multimodality_audio,multimodality_video,
|
134 |
-
licence_name+licence_url, languages, release_date, open_weight
|
135 |
-
|
136 |
-
RR
|
137 |
-
model_name, input_price, output_price,
|
138 |
-
source, release_date
|
139 |
-
|
140 |
-
"""
|
141 |
-
|
142 |
-
|
|
|
1 |
import pandas as pd
|
2 |
import gradio as gr
|
3 |
import os
|
|
|
4 |
|
5 |
+
llm_calc_app = gr.Blocks()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
with llm_calc_app:
|
7 |
|
|
|
|
|
8 |
##################################################
|
|
|
9 |
with gr.Row():
|
10 |
|
11 |
#####################################
|
|
|
14 |
## Language Select
|
15 |
with gr.Column():
|
16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
with gr.Row():
|
18 |
start_date = gr.DateTime(
|
19 |
)
|
|
|
26 |
llm_calc_app.queue()
|
27 |
llm_calc_app.launch()
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|