Spaces:
Running
Running
Update results.py
Browse files- results.py +45 -0
results.py
CHANGED
|
@@ -342,6 +342,45 @@ fig.update_layout(
|
|
| 342 |
# Show figure
|
| 343 |
graph6 = fig
|
| 344 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 345 |
|
| 346 |
intro_div = Div(
|
| 347 |
H2("Perplexity Evaluation on Duplicate Data"),
|
|
@@ -396,26 +435,32 @@ llama_div = Div(
|
|
| 396 |
Section(
|
| 397 |
H2("Llama 3.1 8B"),
|
| 398 |
P("For comparison purpose, we run the same perplexity evaluation with llama 3.1 8B model.")
|
|
|
|
| 399 |
),
|
| 400 |
Section(
|
| 401 |
H3("Perplexity vs Buckets"),
|
| 402 |
Img(src="images/perp-across-diff-buckets-global.png", height = "300", width = "600" ),
|
|
|
|
| 403 |
),
|
| 404 |
Section(
|
| 405 |
H3("Perplexity vs Years"),
|
| 406 |
Img(src="images/prep-across-diff-years-global.png", height = "300", width = "600" ),
|
|
|
|
| 407 |
),
|
| 408 |
Section(
|
| 409 |
H3("Perplexity vs Dump Duplication"),
|
| 410 |
Img(src="images/prep-vs-dump-dup-global.png", height = "300", width = "600" ),
|
|
|
|
| 411 |
),
|
| 412 |
Section(
|
| 413 |
H3("Perplexity vs Local Buckets"),
|
| 414 |
Img(src="images/prep-diff-buckets-local.png", height = "300", width = "600" ),
|
|
|
|
| 415 |
),
|
| 416 |
Section(
|
| 417 |
H3("Perplexity vs Local Dump Duplication"),
|
| 418 |
Img(src="images/prep-vs-dump-dup-global.png", height = "300", width = "600" ),
|
|
|
|
| 419 |
),
|
| 420 |
)
|
| 421 |
|
|
|
|
| 342 |
# Show figure
|
| 343 |
graph6 = fig
|
| 344 |
|
| 345 |
+
#llama graph 1
|
| 346 |
+
|
| 347 |
+
import plotly.graph_objects as go
|
| 348 |
+
|
| 349 |
+
# Data for different years
|
| 350 |
+
data = {
|
| 351 |
+
"2014": [10.036774097041135, 9.306693996275795, 9.442327622499175, 9.015408185880002, 9.94583162297666, 9.141712571508352],
|
| 352 |
+
"2015": [9.46310273785878, 8.763464863196129, 9.075851726027564, 8.868392446242012, 9.113560631617027, 8.037411460181893],
|
| 353 |
+
"2016": [9.41413706166537, 8.645126825996691, 9.527148465147846, 9.120345162203675, 8.9228845723255, 9.14052983061081],
|
| 354 |
+
"2017": [9.50318602661455, 9.473904977192573, 9.755998086072951, 8.968012141869462, 8.895860780054043, 8.757970647106037],
|
| 355 |
+
"2018": [9.007669062339426, 10.95829859145081, 10.128151243953157, 9.451949410987668, 8.863879736723902, 8.440366034517687],
|
| 356 |
+
"2019": [8.388255660116407, 10.676105294328789, 9.728353939624842, 9.381837094065533, 8.401723232809463, 7.5705604983353325],
|
| 357 |
+
"2020": [10.112246017864624, 10.255251179892559, 9.233548505479437, 9.25131862646364, 8.458532176757009, 7.4808205167223525],
|
| 358 |
+
"2021": [10.239269162661959, 9.54987953569235, 9.067380903629866, 9.014261939731549, 8.14345667720481, 7.312019290288715],
|
| 359 |
+
"2022": [9.931951075969451, 9.12737570591033, 8.995868137602248, 9.00805668763514, 7.882044010499616, 7.538858258386088],
|
| 360 |
+
"2023": [8.646614152066428, 8.806922449908505, 8.816629232137835, 8.995152677487027, 7.737747701620713, 6.77703951001925]
|
| 361 |
+
}
|
| 362 |
+
|
| 363 |
+
# Buckets
|
| 364 |
+
buckets = ["1-1", "2-5", "6-10", "11-100", "101-1000", "1001-30000000"]
|
| 365 |
+
|
| 366 |
+
# Create the figure
|
| 367 |
+
fig = go.Figure()
|
| 368 |
+
|
| 369 |
+
# Add lines for each year
|
| 370 |
+
for year, perplexities in data.items():
|
| 371 |
+
fig.add_trace(go.Scatter(x=buckets, y=perplexities, mode='lines+markers', name=year))
|
| 372 |
+
|
| 373 |
+
# Update layout
|
| 374 |
+
fig.update_layout(
|
| 375 |
+
title="Perplexity Across Different Buckets (Global)",
|
| 376 |
+
xaxis_title="Bucket (duplicate count range)",
|
| 377 |
+
yaxis_title="Average Perplexity",
|
| 378 |
+
legend_title="Year"
|
| 379 |
+
)
|
| 380 |
+
|
| 381 |
+
# Show the figure
|
| 382 |
+
llama_graph1 = fig
|
| 383 |
+
|
| 384 |
|
| 385 |
intro_div = Div(
|
| 386 |
H2("Perplexity Evaluation on Duplicate Data"),
|
|
|
|
| 435 |
Section(
|
| 436 |
H2("Llama 3.1 8B"),
|
| 437 |
P("For comparison purpose, we run the same perplexity evaluation with llama 3.1 8B model.")
|
| 438 |
+
plotly2fasthtml(llama_graph1),
|
| 439 |
),
|
| 440 |
Section(
|
| 441 |
H3("Perplexity vs Buckets"),
|
| 442 |
Img(src="images/perp-across-diff-buckets-global.png", height = "300", width = "600" ),
|
| 443 |
+
#plotly2fasthtml(llama_graph1),
|
| 444 |
),
|
| 445 |
Section(
|
| 446 |
H3("Perplexity vs Years"),
|
| 447 |
Img(src="images/prep-across-diff-years-global.png", height = "300", width = "600" ),
|
| 448 |
+
#plotly2fasthtml(llama_graph1),
|
| 449 |
),
|
| 450 |
Section(
|
| 451 |
H3("Perplexity vs Dump Duplication"),
|
| 452 |
Img(src="images/prep-vs-dump-dup-global.png", height = "300", width = "600" ),
|
| 453 |
+
#plotly2fasthtml(llama_graph1),
|
| 454 |
),
|
| 455 |
Section(
|
| 456 |
H3("Perplexity vs Local Buckets"),
|
| 457 |
Img(src="images/prep-diff-buckets-local.png", height = "300", width = "600" ),
|
| 458 |
+
#plotly2fasthtml(llama_graph1),
|
| 459 |
),
|
| 460 |
Section(
|
| 461 |
H3("Perplexity vs Local Dump Duplication"),
|
| 462 |
Img(src="images/prep-vs-dump-dup-global.png", height = "300", width = "600" ),
|
| 463 |
+
#plotly2fasthtml(llama_graph1),
|
| 464 |
),
|
| 465 |
)
|
| 466 |
|