Page MenuHomec4science

dezanet (Sandro Ivo Sebastiano De Zanet)

Projects

Repositories (1)

User Details

User Since
Aug 12 2016, 10:16 (460 w, 2 d)
Availability
Available
Organization
epfl.ch (university)

Recent Activity

Feb 14 2017

dezanet committed R964:87ed85d0bc59: changes for final segmentation (authored by dezanet).
changes for final segmentation
Feb 14 2017, 14:42

Feb 8 2017

dezanet committed R964:5cc87cd5ed4d: Revert "missing reel" (authored by dezanet).
Revert "missing reel"
Feb 8 2017, 14:50
dezanet committed R964:d8dc8e5f7d33: missing reel (authored by dezanet).
missing reel
Feb 8 2017, 14:50
dezanet added a reverting change for R964:d8dc8e5f7d33: missing reel: R964:5cc87cd5ed4d: Revert "missing reel".
Feb 8 2017, 14:50

Feb 7 2017

dezanet committed R964:e5efddc0551e: correct masking (authored by dezanet).
correct masking
Feb 7 2017, 15:06
dezanet committed R964:b3f73b0cdb53: wrong mask.. (authored by dezanet).
wrong mask..
Feb 7 2017, 14:19

Feb 6 2017

dezanet committed R964:5eb1bdecb07c: removed matplotlib (authored by dezanet).
removed matplotlib
Feb 6 2017, 16:40
dezanet committed R964:ccbbedac5e80: in disparity cross link to last layer from input (authored by dezanet).
in disparity cross link to last layer from input
Feb 6 2017, 16:17
dezanet committed R964:f162534c0cf1: fixed test and unet (authored by dezanet).
fixed test and unet
Feb 6 2017, 14:43
dezanet committed R964:23b620bfe440: fixed testing (authored by dezanet).
fixed testing
Feb 6 2017, 13:57
dezanet committed R964:6e162afc442a: testing now segments everything (authored by dezanet).
testing now segments everything
Feb 6 2017, 13:34

Feb 1 2017

dezanet committed R964:6e19ce1ee3e9: wrong part divided by 255 (authored by dezanet).
wrong part divided by 255
Feb 1 2017, 16:03
dezanet committed R964:98ac55961679: fixed bugs (authored by dezanet).
fixed bugs
Feb 1 2017, 15:57
dezanet committed R964:dd8bdb71eca3: fixed typo (authored by dezanet).
fixed typo
Feb 1 2017, 15:35
dezanet committed R964:6c2f431aaaae: masked loss (authored by dezanet).
masked loss
Feb 1 2017, 15:33
dezanet committed R964:fb694d0eab72: reduced kernel size (authored by dezanet).
reduced kernel size
Feb 1 2017, 12:11
dezanet committed R964:f7976129a669: modified layout (authored by dezanet).
modified layout
Feb 1 2017, 12:08
dezanet committed R964:9692516f63eb: bette output (authored by dezanet).
bette output
Feb 1 2017, 11:52
dezanet committed R964:24861ee8e782: removed regulariser on dilation/erosion (authored by dezanet).
removed regulariser on dilation/erosion
Feb 1 2017, 11:47

Jan 31 2017

dezanet committed R964:8b9a53fa62d5: larger scale (authored by dezanet).
larger scale
Jan 31 2017, 13:32
dezanet committed R964:d8df2c9cf8a3: regs (authored by dezanet).
regs
Jan 31 2017, 12:17
dezanet committed R964:8a1a2f90b40f: dilation and erosion in different places (authored by dezanet).
dilation and erosion in different places
Jan 31 2017, 12:08
dezanet committed R964:dedc3031c6db: erosion layer (authored by dezanet).
erosion layer
Jan 31 2017, 11:32
dezanet committed R964:691854a501c3: added nonlinearity (authored by dezanet).
added nonlinearity
Jan 31 2017, 10:43
dezanet committed R964:d64629ead64f: added linear combination (authored by dezanet).
added linear combination
Jan 31 2017, 10:41
dezanet committed R964:bbe4f88c5185: larger dilation (authored by dezanet).
larger dilation
Jan 31 2017, 10:39
dezanet committed R964:34bedf953c14: moved dilation (authored by dezanet).
moved dilation
Jan 31 2017, 10:37

Jan 30 2017

dezanet committed R964:8140b77fcbbd: less regularisation (authored by dezanet).
less regularisation
Jan 30 2017, 16:40
dezanet committed R964:b60251ef4039: convolution and dilation (authored by dezanet).
convolution and dilation
Jan 30 2017, 16:37
dezanet committed R964:d19731d87f9b: umorphnet (authored by dezanet).
umorphnet
Jan 30 2017, 16:26
dezanet committed R964:6408345e128c: morphology (authored by dezanet).
morphology
Jan 30 2017, 16:24
dezanet committed R964:ed11c9ede454: he_normal initialisation, RELU instead of ELU (authored by dezanet).
he_normal initialisation, RELU instead of ELU
Jan 30 2017, 14:31
dezanet committed R964:0fa456f74b0a: removed augmentation (authored by dezanet).
removed augmentation
Jan 30 2017, 13:17
dezanet committed R964:e712783e709d: added mirror padding (authored by dezanet).
added mirror padding
Jan 30 2017, 13:11
dezanet committed R964:c149e6e3a387: valid convolution, l2 regularisation (authored by dezanet).
valid convolution, l2 regularisation
Jan 30 2017, 13:09

Jan 19 2017

dezanet committed R964:d30568b67f87: add batch normalisation (authored by dezanet).
add batch normalisation
Jan 19 2017, 12:18
dezanet committed R964:1661a4b01cf3: add rotation again (authored by dezanet).
add rotation again
Jan 19 2017, 12:01
dezanet committed R964:27821e2810a0: back to original norm (authored by dezanet).
back to original norm
Jan 19 2017, 11:41
dezanet committed R964:17934843cf08: no augmentation (authored by dezanet).
no augmentation
Jan 19 2017, 11:37
dezanet committed R964:28d3787b3ff1: different loss (authored by dezanet).
different loss
Jan 19 2017, 11:29
dezanet committed R964:3cbe32242a3f: new data normalization (authored by dezanet).
new data normalization
Jan 19 2017, 11:09
dezanet committed R964:4faccc37c63e: add rotation augmentation (authored by dezanet).
add rotation augmentation
Jan 19 2017, 09:54

Jan 12 2017

dezanet committed R964:a5ee7fa0d36b: faster activation (authored by dezanet).
faster activation
Jan 12 2017, 17:15
dezanet committed R964:46cb8ef0e096: reunite (authored by dezanet).
reunite
Jan 12 2017, 15:36

Jan 10 2017

dezanet committed R964:18bb39af0995: back to net (authored by dezanet).
back to net
Jan 10 2017, 10:59

Jan 9 2017

dezanet committed R964:732385cd4761: more jumping (authored by dezanet).
more jumping
Jan 9 2017, 16:18
dezanet committed R964:95212031368b: more filters, batchsize10 (authored by dezanet).
more filters, batchsize10
Jan 9 2017, 13:24
dezanet committed R964:e20f0871e9df: merge (authored by dezanet).
merge
Jan 9 2017, 12:30
dezanet committed R964:fdabe7c466e9: double filters (authored by dezanet).
double filters
Jan 9 2017, 12:13
dezanet committed R964:be53ac9f93e7: dilated test (authored by dezanet).
dilated test
Jan 9 2017, 11:59

Dec 22 2016

dezanet changed the visibility for lineseg.
Dec 22 2016, 13:36
dezanet added a watcher for lineseg: dezanet.
Dec 22 2016, 13:35
dezanet committed R964:ae283546c00b: removed reconstruction (authored by dezanet).
removed reconstruction
Dec 22 2016, 12:11
dezanet committed R964:b437cb13c929: last stand (authored by dezanet).
last stand
Dec 22 2016, 12:10

Dec 5 2016

dezanet committed R964:5c250b837663: fixed callback (authored by dezanet).
fixed callback
Dec 5 2016, 14:42
dezanet committed R964:d867b1af02e4: flow net needs two outputs (authored by dezanet).
flow net needs two outputs
Dec 5 2016, 14:30
dezanet committed R964:584822c13c2d: more verbose output (authored by dezanet).
more verbose output
Dec 5 2016, 14:19
dezanet committed R964:33bd35298d07: fixing problems with disparity regression (authored by dezanet).
fixing problems with disparity regression
Dec 5 2016, 14:19

Nov 30 2016

dezanet committed R964:94edbec404a4: wrong types (authored by dezanet).
wrong types
Nov 30 2016, 13:25
dezanet committed R964:fd3f888a6be3: domain adaption with new model (authored by dezanet).
domain adaption with new model
Nov 30 2016, 13:17

Nov 29 2016

dezanet committed R964:8ffee70ce61c: use float images (authored by dezanet).
use float images
Nov 29 2016, 15:12
dezanet committed R964:c192cda5673b: moar filters (authored by dezanet).
moar filters
Nov 29 2016, 15:11
dezanet committed R964:c7654b962f35: moved normalise to data.py (authored by dezanet).
moved normalise to data.py
Nov 29 2016, 11:58
dezanet committed R964:8a7a2cf0f379: normalization in test (authored by dezanet).
normalization in test
Nov 29 2016, 11:55
dezanet committed R964:6fbedcdf73bd: double width (authored by dezanet).
double width
Nov 29 2016, 11:05
dezanet committed R964:e8c5878f03a4: use loss (authored by dezanet).
use loss
Nov 29 2016, 11:00
dezanet committed R964:a9bb0e588967: more conv per level (authored by dezanet).
more conv per level
Nov 29 2016, 10:52
dezanet committed R964:6b22507eff9f: reduce k (authored by dezanet).
reduce k
Nov 29 2016, 10:46
dezanet committed R964:72c6ee204504: make data great again (authored by dezanet).
make data great again
Nov 29 2016, 10:42
dezanet committed R964:e0efb5ec202d: augmentation failed (authored by dezanet).
augmentation failed
Nov 29 2016, 10:38
dezanet committed R964:e2d6c8953022: augmentation (authored by dezanet).
augmentation
Nov 29 2016, 10:33
dezanet committed R964:e88adfc77a2b: increased learning rate (authored by dezanet).
increased learning rate
Nov 29 2016, 10:29
dezanet committed R964:c82e40047b48: return correct sized image (authored by dezanet).
return correct sized image
Nov 29 2016, 10:27
dezanet committed R964:0ba59ba4b315: moving the data to -100,100 (authored by dezanet).
moving the data to -100,100
Nov 29 2016, 10:06

Nov 28 2016

dezanet committed R964:ad4e6393599f: no normalisation (authored by dezanet).
no normalisation
Nov 28 2016, 17:04
dezanet committed R964:13f2062cb4d3: no ReLU in first layer (authored by dezanet).
no ReLU in first layer
Nov 28 2016, 16:21
dezanet committed R964:60fe0433734d: corrected normalisation (authored by dezanet).
corrected normalisation
Nov 28 2016, 15:58
dezanet committed R964:6343f8e1a0a4: softmax (authored by dezanet).
softmax
Nov 28 2016, 15:53
dezanet committed R964:314b613f9644: different normalisation (authored by dezanet).
different normalisation
Nov 28 2016, 15:26
dezanet committed R964:cd22810abef6: reduced batch size (authored by dezanet).
reduced batch size
Nov 28 2016, 15:12
dezanet committed R964:01fc096a601a: normalisation bug (authored by dezanet).
normalisation bug
Nov 28 2016, 15:02
dezanet committed R964:e70cf9c4ea39: try normalisation (authored by dezanet).
try normalisation
Nov 28 2016, 15:00
dezanet committed R964:a59f301680d1: slower learning (authored by dezanet).
slower learning
Nov 28 2016, 14:51
dezanet committed R964:53be0d61f1c1: no normalisation (authored by dezanet).
no normalisation
Nov 28 2016, 14:45
dezanet committed R964:bf61b63df141: small cleanup (authored by dezanet).
small cleanup
Nov 28 2016, 14:40
dezanet committed R964:1da354ec59d6: remove generator (authored by dezanet).
remove generator
Nov 28 2016, 14:30
dezanet committed R964:8ab6537f0f89: included batch size (authored by dezanet).
included batch size
Nov 28 2016, 14:14
dezanet committed R964:4bd57a969d57: callback images generation (authored by dezanet).
callback images generation
Nov 28 2016, 13:56
dezanet committed R964:a58311225bc7: no augmentation (authored by dezanet).
no augmentation
Nov 28 2016, 13:14
dezanet committed R964:b594d30c9ec6: only flip (authored by dezanet).
only flip
Nov 28 2016, 12:03
dezanet committed R964:f2c75ed5782a: increased number of samples (authored by dezanet).
increased number of samples
Nov 28 2016, 11:52
dezanet committed R964:4841838b8b00: reduced amount of augmentation (authored by dezanet).
reduced amount of augmentation
Nov 28 2016, 11:50
dezanet committed R964:3fa3a03b4310: proper augmentation (authored by dezanet).
proper augmentation
Nov 28 2016, 11:44

Nov 25 2016

dezanet committed R964:cb46a795235d: back to unit (authored by dezanet).
back to unit
Nov 25 2016, 11:11

Nov 22 2016

dezanet committed R964:e90cd9649b5e: increased recursion limit (authored by dezanet).
increased recursion limit
Nov 22 2016, 16:05
dezanet committed R964:dc9073ad010a: naming correction (authored by dezanet).
naming correction
Nov 22 2016, 14:11
dezanet committed R964:5a1e4c3153a9: more convolutions (authored by dezanet).
more convolutions
Nov 22 2016, 14:09
dezanet committed R964:3f644ffa7187: batchsize=1 (authored by dezanet).
batchsize=1
Nov 22 2016, 13:04
dezanet committed R964:63e3da027ff8: shallower (authored by dezanet).
shallower
Nov 22 2016, 12:56
dezanet committed R964:aa982344fe50: no identity blocks (authored by dezanet).
no identity blocks
Nov 22 2016, 12:50