Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Aug 1, 2023

• update config

  nbtin

  committed
  b4815cd

• update config

  nbtin

  committed
  11c673d

• update config

  nbtin

  committed
  ad8db39

• add prune model method

  nbtin

  committed
  07f1583

• update counting time in test file

  nbtin

  committed
  698db18

• update test

  nbtin

  committed
  bc9d45d

• update test

  nbtin

  committed
  9c258d1

• update test

  nbtin

  committed
  4bb14b1

• update eval

  nbtin

  committed
  633c694

• update eval

  nbtin

  committed
  5bdf9f2

• Add progress_bar customization.

  nbtin

  committed
  dd532fc

Commits on Mar 27, 2022

• docs: fix table format

  weiaicunzai

  committed
  11d8418

Commits on Dec 29, 2020

• feat: print epoch during testing

  weiaicunzai

  committed
  2149cb5

• fix: google large gpu memory usage

  weiaicunzai

  committed
  49ddb26

• feat: change argument gpu to store_true

  weiaicunzai

  committed
  b9cca00

• feat: print gpu info only once when testing

  weiaicunzai

  committed
  0df7238

Commits on Dec 28, 2020

• fix: large gpu memory required by googlenet, add more downsampling

  weiaicunzai

  committed
  a9c8225

• refactor: change output to x in googlenet forward function

  weiaicunzai

  committed
  5bd0fa8

• refactor: put lr.step() after optimizer.step()

  weiaicunzai

  committed
  ca235c2

• refactor: add runs and checkpoints folders to .gitignore file

  weiaicunzai

  committed
  ffdb814

• dcos: remove gpu memory util, memory util is effected by input image

  Show description for ddf1db4
  weiaicunzai

  committed
  ddf1db4

• docs: add sotchastic depth 101 result

  weiaicunzai

  committed
  1c7ab20

• docs: add stochasticdepth50 result

  weiaicunzai

  committed
  01781f9

• docs: add stochasticdepth34 result

  weiaicunzai

  committed
  1d96935

• feat: put tensor p inside a function instead of class member

  weiaicunzai

  committed
  bf2fa9d

• fix: always return false due to different tensor shape in survial method

  weiaicunzai

  committed
  ca25ee9

• fix: set torch.ones(1) on the same device as the output of torch.bernoulli()

  weiaicunzai

  committed
  5974f22

• fix: remove redundant newline

  weiaicunzai

  committed
  e471d6a

• fix: torch.jit.tracke error: change block type to torch.jit.ScriptMedule

  Show description for 60b7236
  weiaicunzai

  committed
  60b7236

• refactor: use pytorch buildin bernoulli distribution sampler

  weiaicunzai

  committed
  4cc6a35

• feat: add stochastic depth network

  weiaicunzai

  committed
  ba50283

• feat: push for test

  weiaicunzai

  committed
  aecf3af

• feat: put tensor p inside a function instead of class member

  weiaicunzai

  committed
  eaa4090

• feat: print gpu usage

  weiaicunzai

  committed
  979c8ab

• fix: always return false due to different tensor shape in survial method

  weiaicunzai

  committed
  09bc55e

Pagination

PreviousNext