Skip to content

Instantly share code, notes, and snippets.

View terrychenism's full-sized avatar

Terry Chen terrychenism

210 followers · 12 following
  • U.S.
View GitHub Profile
@terrychenism
terrychenism / sort.py
Last active March 12, 2019 07:11
simple baseline of mot
from __future__ import print_function
import os.path
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as patches
from skimage import io
from sklearn.utils.linear_assignment_ import linear_assignment
import glob
import time
import argparse
@terrychenism
terrychenism / resnext-152.log
Created December 28, 2016 04:34
This file has been truncated, but you can view the full file.
INFO:root:Namespace(aug_level=3, batch_size=256, bn_mom=0.9, data_dir='/home/ubuntu/imagenet', data_type='imagenet', depth=152, frequent=50, gpus='0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15', kv_store='device', list_dir='./', lr=0.1, memonger=False, model_load_epoch=0, mom=0.9, num_classes=1000, num_examples=1281167, retrain=False, wd=0.0001, workspace=512)
[07:24:40] src/io/iter_image_recordio.cc:221: ImageRecordIOParser: /home/ubuntu/imagenet/train_480_q90.rec, use 4 threads for decoding..
[07:24:43] src/io/iter_image_recordio.cc:221: ImageRecordIOParser: /home/ubuntu/imagenet/val_256_q90.rec, use 4 threads for decoding..
INFO:root:Start training with [gpu(0), gpu(1), gpu(2), gpu(3), gpu(4), gpu(5), gpu(6), gpu(7), gpu(8), gpu(9), gpu(10), gpu(11), gpu(12), gpu(13), gpu(14), gpu(15)]
INFO:root:Epoch[0] Batch [50] Speed: 122.55 samples/sec Train-accuracy=0.001797
INFO:root:Epoch[0] Batch [50] Speed: 122.55 samples/sec Train-top_k_accuracy_5=0.005625
INFO:root:Epoch[0] Batch [100] Speed: 118.30 samples/sec Train-a
@terrychenism
terrychenism / resnext-50.log
Created December 17, 2016 01:13
This file has been truncated, but you can view the full file.
INFO:root:Namespace(aug_level=2, batch_size=256, bn_mom=0.9, data_dir='/H1/tairuic/ResNet/data/imagenet', data_type='imagenet', depth=50, frequent=50, gpus='4,5,6,7', kv_store='device', list_dir='./', lr=0.1, memonger=False, model_load_epoch=0, mom=0.9, num_classes=1000, num_examples=1281167, retrain=False, wd=0.0001, workspace=512)
[23:43:58] src/io/iter_image_recordio.cc:219: ImageRecordIOParser: /H1/tairuic/ResNet/data/imagenet/train_480_q90.rec, use 4 threads for decoding..
[23:43:59] src/io/iter_image_recordio.cc:219: ImageRecordIOParser: /H1/tairuic/ResNet/data/imagenet/val_256_q90.rec, use 4 threads for decoding..
INFO:root:Start training with [gpu(4), gpu(5), gpu(6), gpu(7)]
INFO:root:Epoch[0] Batch [50] Speed: 294.88 samples/sec Train-accuracy=0.001016
INFO:root:Epoch[0] Batch [50] Speed: 294.88 samples/sec Train-top_k_accuracy_5=0.006094
INFO:root:Epoch[0] Batch [100] Speed: 283.82 samples/sec Train-accuracy=0.000859
INFO:root:Epoch[0] Batch [100] Speed: 283.82 samples/sec Train-top_k_accuracy_5=0.0
@terrychenism
terrychenism / resnext-101.log
Created December 7, 2016 02:56
This file has been truncated, but you can view the full file.
INFO:root:Namespace(aug_level=2, batch_size=256, bn_mom=0.9, data_dir='data/imagenet', data_type='imagenet', depth=101, frequent=50, gpus='0,1,2,3,4,5,6,7', kv_store='device', list_dir='./', lr=0.1, memonger=False, model_load_epoch=0, mom=0.9, num_classes=1000, num_examples=1281167, retrain=False, wd=0.0001, workspace=512)
[00:03:29] src/io/iter_image_recordio.cc:221: ImageRecordIOParser: data/imagenet/train_480_q90.rec, use 4 threads for decoding..
[00:03:32] src/io/iter_image_recordio.cc:221: ImageRecordIOParser: data/imagenet/val_256_q90.rec, use 4 threads for decoding..
INFO:root:Start training with [gpu(0), gpu(1), gpu(2), gpu(3), gpu(4), gpu(5), gpu(6), gpu(7)]
INFO:root:Epoch[0] Batch [50] Speed: 136.63 samples/sec Train-accuracy=0.001328
INFO:root:Epoch[0] Batch [50] Speed: 136.63 samples/sec Train-top_k_accuracy_5=0.006406
INFO:root:Epoch[0] Batch [100] Speed: 132.44 samples/sec Train-accuracy=0.001328
INFO:root:Epoch[0] Batch [100] Speed: 132.44 samples/sec Train-top_k_accuracy_5=0.005625
INFO:root: