Skip to content

Instantly share code, notes, and snippets.

View zer0n's full-sized avatar

Kenneth Tran zer0n

185 followers · 15 following
View GitHub Profile
@zer0n
zer0n / pytorch_hybrid.ipynb
Created February 9, 2019 05:24
LEARNING HYBRID FRONTEND SYNTAX THROUGH EXAMPLE
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@zer0n
zer0n / main.personal.py
Last active May 8, 2018 22:43
Example personal main file
from train import _get_parser as get_trainer_parser, train
# Get default arguments. Then you can change the default values for a few arguments
args = get_trainer_parser().parse_args([])
args.verbose = True
train(args)
@zer0n
zer0n / CNN Tutorial on Text Classification.ipynb
Created March 1, 2017 01:22
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@zer0n
zer0n / CNN Tutorial on MNIST.ipynb
Created March 1, 2017 01:17
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@zer0n
zer0n / reclaimWindows10.ps1
Created January 8, 2017 00:45 — forked from alirobe/reclaimWindows10.ps1
"Reclaim Windows 10" turns off a bunch of unnecessary Windows 10 telemetery, removes bloatware, and privacy invasions. Review and tweak before running. Scripts for reversing are included and commented. Fork via https://github.com/Disassembler0 (different defaults)
##########
# Win10 Initial Setup Script
# Author: Disassembler <disassembler@dasm.cz>
# Version: 1.7, 2016-08-15
# dasm's script: https://github.com/Disassembler0/Win10-Initial-Setup-Script/
# THIS IS A PERSONALIZED VERSION
# This script leaves more MS defaults on, including MS security features.
# Tweaked based on personal preferences for @alirobe 2016-11-16 - v1.7.1
@zer0n
zer0n / rfcn_alt_opt_5step_ohem_rpn4_ResNet-101_.txt.2016-11-06_23-33-58
Created November 8, 2016 02:33
+ echo Logging output to experiments/logs/rfcn_alt_opt_5step_ohem_rpn4_ResNet-101_.txt.2016-11-06_23-33-58
Logging output to experiments/logs/rfcn_alt_opt_5step_ohem_rpn4_ResNet-101_.txt.2016-11-06_23-33-58
+ ./tools/train_rfcn_alt_opt_5stage.py --gpu 4 --net_name ResNet-101 --weights data/imagenet_models/ResNet-101-model.caffemodel --imdb coco_2014_train --imdb_test coco_2014_minival --cfg experiments/cfgs/rfcn_alt_opt_5step_ohem.yml --model rfcn_alt_opt_5step_ohem
Called with args:
Namespace(cfg_file='experiments/cfgs/rfcn_alt_opt_5step_ohem.yml', gpu_id=4, imdb_name='coco_2014_train', imdb_test_name='coco_2014_minival', model_name='rfcn_alt_opt_5step_ohem', net_name='ResNet-101', pretrained_model='data/imagenet_models/ResNet-101-model.caffemodel', set_cfgs=None)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Stage 0 RPN, compute normalization means and stds
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
loading annotations into memory...
Done (t=11.90s)
@zer0n
zer0n / rfcn_end2end_ResNet-101_.txt.2016-11-07_01-23-01
Created November 8, 2016 02:32
+ echo Logging output to experiments/logs/rfcn_end2end_ResNet-101_.txt.2016-11-07_01-23-01
Logging output to experiments/logs/rfcn_end2end_ResNet-101_.txt.2016-11-07_01-23-01
+ ./tools/train_net.py --gpu 4 --solver models/coco/ResNet-101/rfcn_end2end/solver_ohem.prototxt --weights data/imagenet_models/ResNet-101-model.caffemodel --imdb coco_2014_train --iters 490000 --cfg experiments/cfgs/rfcn_end2end_ohem.yml
Called with args:
Namespace(cfg_file='experiments/cfgs/rfcn_end2end_ohem.yml', gpu_id=4, imdb_name='coco_2014_train', max_iters=490000, pretrained_model='data/imagenet_models/ResNet-101-model.caffemodel', randomize=False, set_cfgs=None, solver='models/coco/ResNet-101/rfcn_end2end/solver_ohem.prototxt')
Using config:
{'DATA_DIR': '/home/zer0ne/src/R-FCN/data',
'DEDUP_BOXES': 0.0625,
'EPS': 1e-14,
'EXP_DIR': 'rfcn_end2end_ohem',
@zer0n
zer0n / Instagram.lst
Created March 25, 2016 08:17
http://i.ytimg.com/vi/rGAqBy-D8wc/maxresdefault.jpg
http://blog.defgrip.net/wp-content/uploads/2013/08/roof1.jpg
https://scontent.cdninstagram.com/hphotos-xpf1/t51.2885-15/s320x320/e15/10808825_409571005859448_326802933_n.jpg
http://www.madpac.nl/wp-content/uploads/2015/12/Chris2-1024x684.jpg
http://thehundreds.com/wp-content/uploads/2015/01/h-big-tatt-3.jpg
http://scontent-a.cdninstagram.com/hphotos-xfp1/t51.2885-15/10375657_582076715241103_588595470_a.jpg
http://smpfreeskiercom.c.presscdn.com/wp-content/uploads/2015/09/Erik-Roner.jpg
http://41.media.tumblr.com/7f20e204983f9cef0172edc464eaebdd/tumblr_mnmyl1lxtF1qienb7o1_1280.jpg
http://www.geeksofdoom.com/GoD/img/2013/10/2013-10-12-nycc-walking-dead-steven-yeun-11.jpg
http://media-cache-ec0.pinimg.com/736x/bd/f6/58/bdf6583cbbf839aea52bc5782a70d474.jpg
@zer0n
zer0n / Visual Genome.ipynb
Last active February 2, 2016 07:36
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@zer0n
zer0n / theano_benchmark.py
Last active December 14, 2015 17:31
import os
os.environ['THEANO_FLAGS'] = 'mode=FAST_RUN,device=gpu,floatX=float32,nvcc.fastmath=True'
import theano
import time
import numpy as np
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD