Bbbbbnnn/ai-phone-leaderboard

fix: filter out submissions with ts or pp less than 0.01

chore: filter out anomalies (z_threshold=9.0)

fix: for Device Duel get max for each device/model combination

chore: minor tab title correction

fix: replace depricated titlefont with title_font

chore: order charts in Device Duel Arena

feat: refactor Device Duel

chore: clean up device ranking tab

chore: minor updates

feat: display top 30 devices

feat: add plot ranking ladder

chore: add caching for data load

chore: add caching for glicko calculation

chore: remove platform ranking

chore: cleanup old scoring code

feat: add Glicko2 ranking

feat: use android device marketing name

add ios device name sanitation

fix: expand Device width

chore(ui): increase ranking table height

chore: update doc

feat: add quantization to "Model Size vs Performance" plot

feat: add TG and PP scores

feat: re order size category for "Rankings by Model Size"

pre select plot for Model Size vs Performance based on ranking

feat: add Model Size vs Performance

fix(ui): rename Quantization Level from factor to quantization code

feat: replace index with ranking in the tables

feat(scoring): use model size as direct multiplier

fix(ranking): use normalized device id for aggregation

chore: default pp tg to 512 and 128 for the plots

refactor(quant factors): linearize the factors

fix: import st

refactor(scoring): Filter benchmarks to standard conditions only

feat: add ranking

fix column_order

Track binary files with Git LFS

add gpu layers

update guide col

add guide

align the header

update header

add guide to submit

rename memory usage to peak memory

fix plots

add Versions

fix filters

refactor filters

apply header

update table and plots