Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 22, 2019

• run_servers_and_cl.py: redirect the stderr of run_servers to a log file

  francisyyan

  committed
  0855c21

  Copy full SHA for 0855c21

• ttp.py: add argument --cl to perform continual learning

  francisyyan

  committed
  b692772

  Copy full SHA for b692772

• run_servers_and_cl.py: find a non-existent model directory to save

  francisyyan

  committed
  c85e18b

  Copy full SHA for c85e18b

Commits on Jan 21, 2019

• plot_ssim_rebuffer.py: exclude startup delay from rebuffer time

  francisyyan

  committed
  41ad0cd

  Copy full SHA for 41ad0cd

• Finish run_servers_and_cl.py; ttp.py to be completed with "--cl"

  francisyyan

  committed
  dc066b8

  Copy full SHA for dc066b8

• Add skeleton of run_servers_and_cl.py: performs continual learning at 11AM (UTC) every day

  francisyyan

  committed
  066831d

  Copy full SHA for 066831d

Commits on Jan 20, 2019

• Rename train_ttp.py to ttp.py and add predict

  francisyyan

  committed
  63a1127

  Copy full SHA for 63a1127

Commits on Jan 18, 2019

• train_ttp.py: save checkpoints every 50 epochs

  francisyyan

  committed
  42dc99a

  Copy full SHA for 42dc99a

• Tune the coefficients for ssim db

  NeverlMole

  committed
  55a6d78

  Copy full SHA for 55a6d78

• Set the max and min of chunk ssim values

  NeverlMole

  committed
  0384162

  Copy full SHA for 0384162

• Ignore ssim difference loss for the first chunk

  NeverlMole

  committed
  1a42955

  Copy full SHA for 1a42955

• Use SSIM dB in QoE objectives

  francisyyan

  committed
  af8ee27

  Copy full SHA for af8ee27

Commits on Jan 17, 2019

• index.html: revise again

  keithw

  committed
  070fe82

  Copy full SHA for 070fe82

• index.html: revise

  keithw

  committed
  12c3da2

  Copy full SHA for 12c3da2

• Update front page to improve "landing page experience"

  keithw

  committed
  c271c42

  Copy full SHA for c271c42

Commits on Jan 16, 2019

• Adjust language on timeout error message

  keithw

  committed
  9e58c91

  Copy full SHA for 9e58c91

• Modify Reddit conversion pixel

  keithw

  committed
  c2e60f1

  Copy full SHA for c2e60f1

• Set the robust_mpc explicitly

  NeverlMole

  committed
  820eed9

  Copy full SHA for 820eed9

• Add robust_mpc ABR

  NeverlMole

  committed
  1b89041

  Copy full SHA for 1b89041

• train_ttp.py: fix congestion control filter

  francisyyan

  committed
  8ec1dcd

  Copy full SHA for 8ec1dcd

Commits on Jan 15, 2019

• train_ttp.py: tiny bug fix

  francisyyan

  committed
  3e181b8

  Copy full SHA for 3e181b8

• train_ttp.py: filter input data by congestion control

  francisyyan

  committed
  92c1bec

  Copy full SHA for 92c1bec

• Move function to retrieve experimental configs to helpers.py

  francisyyan

  committed
  69b502f

  Copy full SHA for 69b502f

• Update hyperparameters used in the currently deployed model

  francisyyan

  committed
  d60e628

  Copy full SHA for d60e628

• Amend the order and scale of the prepared inputs to ttp

  NeverlMole

  committed
  6b6c98d

  Copy full SHA for 6b6c98d

• Move static constants inside class; use .value() to check optional

  francisyyan

  committed
  d044f80

  Copy full SHA for d044f80

• List secondary dependencies of dynamic libraries explicitly

  francisyyan

  committed
  85a0a11

  Copy full SHA for 85a0a11

Commits on Jan 14, 2019

• Add puffer_ttp ABR

  NeverlMole

  committed
  ef7213b

  Copy full SHA for ef7213b

• Add the handling for all-ban cases

  NeverlMole

  committed
  02e6b3b

  Copy full SHA for 02e6b3b

• Record chunk sizes in puffer ABR

  NeverlMole

  committed
  743f6df

  Copy full SHA for 743f6df

• Best practice of rpath: use $ORIGIN instead of absolute path

  francisyyan

  committed
  bc555b1

  Copy full SHA for bc555b1

Commits on Jan 13, 2019

• Fix build: rpath needs to be absolute

  francisyyan

  committed
  2a52773

  Copy full SHA for 2a52773

• Add Reddit tracking pixel

  keithw

  committed
  adef9c5

  Copy full SHA for adef9c5

• Fix number of threads used in train_ttp.py

  francisyyan

  committed
  96bccea

  Copy full SHA for 96bccea

• Replace/add fetch_submodules.py: fetch release of libtorch built on our own

  francisyyan

  committed
  846d3de

  Copy full SHA for 846d3de

Pagination

Previous

Next