Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Nov 3, 2021

• updated README

  DuncanMays

  committed
  1a95ee4

  Copy full SHA for 1a95ee4

Commits on Oct 29, 2021

• updated readme

  DuncanMays

  committed
  ab747e3

  Copy full SHA for ab747e3

Commits on Oct 27, 2021

• started running elarning jobs locally

  DuncanMays

  committed
  6c2e4cf

  Copy full SHA for 6c2e4cf

Commits on Oct 26, 2021

• got init working

  DuncanMays

  committed
  533f5b4

  Copy full SHA for 533f5b4

Commits on Jul 9, 2021

• plotted some results

  DuncanMays

  committed
  ea4e3ba

  Copy full SHA for ea4e3ba

Commits on Jul 6, 2021

• finished third trial

  DuncanMays

  committed
  549cf16

  Copy full SHA for 549cf16

Commits on Jul 2, 2021

• finished second round of trials

  DuncanMays

  committed
  08f693f

  Copy full SHA for 08f693f

Commits on Jun 30, 2021

• completed first trial for all work functions

  DuncanMays

  committed
  6e83307

  Copy full SHA for 6e83307

• V4 had issues transmitting paramters to workers, fixed the issue by using the shared input paramter, rather than the slice parameter

  DuncanMays

  committed
  48c3d80

  Copy full SHA for 48c3d80

Commits on Jun 29, 2021

• changed main.js to give the training time and work function on completion. Switching testing to the V4 scheduler, highlighting the results collected on V3.

  DuncanMays

  committed
  ed828ff

  Copy full SHA for ed828ff

Commits on Jun 16, 2021

• third of the way through the first trial of the third work functin, commiting before I leave

  DuncanMays

  committed
  84a4f46

  Copy full SHA for 84a4f46

• fixed a memory leak issue that was crashing long-time workers. Completed the first trials for the first two work functions

  DuncanMays

  committed
  fe2c979

  Copy full SHA for fe2c979

• moving code to work

  DuncanMays

  committed
  9c706a6

  Copy full SHA for 9c706a6

Commits on Jun 15, 2021

• pushing spreadsheet skeleton for data collection

  DuncanMays

  committed
  ad6a23b

  Copy full SHA for ad6a23b

Commits on Jun 14, 2021

• workers now report the number of data shards they downloaded

  DuncanMays

  committed
  7a41474

  Copy full SHA for 7a41474

Commits on Jun 11, 2021

• all work functions are operational

  DuncanMays

  committed
  74543d7

  Copy full SHA for 74543d7

Commits on Jun 9, 2021

• its working alright now

  DuncanMays

  committed
  0e5e94d

  Copy full SHA for 0e5e94d

• everything aggregates alright, no crashing, but the final parameters arent very good

  DuncanMays

  committed
  7b777e2

  Copy full SHA for 7b777e2

• A bug was that the second parameter of the work function, shared_input, was undefined. Im fairly certain that this is a bug in DCP. Instead I set shared input to be the first parameter, which comes…

  Show description for e2cab70
  DuncanMays

  committed
  e2cab70

  Copy full SHA for e2cab70

Commits on Jun 8, 2021

• working through some bugs that occur in worker, commiting before switching to another task

  DuncanMays

  committed
  5e30bdf

  Copy full SHA for 5e30bdf

• got some preliminary data loading working

  DuncanMays

  committed
  865e496

  Copy full SHA for 865e496

Commits on May 10, 2021

• got data download working for the most complicated work function, the other methods are derivatives of this one, so we should just be able to copy and paste snippets of code from this one

  DuncanMays

  committed
  659d96e

  Copy full SHA for 659d96e

• cleaning up the repo

  DuncanMays

  committed
  3d19edc

  Copy full SHA for 3d19edc

Commits on Apr 20, 2021

• attempted to implement variable download, still running into bugs

  DuncanMays

  committed
  86ff765

  Copy full SHA for 86ff765

Commits on Apr 4, 2021

• collected some data with the new benchmark, looking good

  DuncanMays

  committed
  0090ce9

  Copy full SHA for 0090ce9

Commits on Apr 2, 2021

• implemented new benchmark that directly measures sample rate

  DuncanMays

  committed
  14cb918

  Copy full SHA for 14cb918

• implemented new benchmark that directly measueres sample rate on randomly generated data

  DuncanMays

  committed
  129040b

  Copy full SHA for 129040b

Commits on Mar 25, 2021

• ran some tests at 30 epochs, less weird results

  DuncanMays

  committed
  1d908d3

  Copy full SHA for 1d908d3

Commits on Mar 24, 2021

• added hashing to the data commit script so that the same data cannot be commited twice

  DuncanMays

  committed
  2b0e71a

  Copy full SHA for 2b0e71a

Commits on Mar 23, 2021

• ran another FLOPS test, changed data recording a little bit

  DuncanMays

  committed
  d55c52c

  Copy full SHA for d55c52c

• implemented 2-phase data recording and added comments

  DuncanMays

  committed
  a735ca0

  Copy full SHA for a735ca0

• uploading some results

  DuncanMays

  committed
  d9ebea7

  Copy full SHA for d9ebea7

Commits on Mar 22, 2021

• ran a partially successful test, got some results and made a plot

  DuncanMays

  committed
  59b5a1c

  Copy full SHA for 59b5a1c

• got plotting working and results are saved when control C is pressed, so we can cancel a long job if it takes too long and still keep what results we have

  DuncanMays

  committed
  ba9ba13

  Copy full SHA for ba9ba13

• FLOPS per sample rate calculation working and recording results successfully, now for a larger scale test

  DuncanMays

  committed
  a956338

  Copy full SHA for a956338