amitkayal
March 17, 2019, 5:18am
1
I am trying to work on Tiny Imagenet data from kaggle and wondering if anyone has done this using fast.ai…The data link is https://www.kaggle.com/c/tiny-imagenet …This dataset seems to be extremely challenging one…
1 Like
adrian
March 18, 2019, 10:44am
2
I did for part 2 2018. May get are least something you can utilise from here:
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Lesson8 pascal object detecion applied to Tiny ImageNet Visual Recognition Challenge\n",
"https://tiny-imagenet.herokuapp.com/\n",
"\n",
"\"*Tiny Imagenet has 200 classes. Each class has 500 training images, 50 validation images, and 50 test images. We have released the training and validation sets with images and annotations. We provide both class labels and bounding boxes as annotations; however, you are asked only to predict the class label of each image without localizing the objects. The test set is released without labels*\"\n",
"\n",
"see also Tiny Imagenet Challenge, Yinbin Ma (2017) http://cs231n.stanford.edu/reports/2017/pdfs/935.pdf as well as http://cs231n.stanford.edu/reports/2017/pdfs/926.pdf for papers using this dataset\n"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"ExecuteTime": {
"end_time": "2018-03-30T13:04:00.727976Z",
show original
1 Like
amitkayal
March 20, 2019, 10:29am
3
thanks and will look into this one…This looks like using old fast.ai libs