Skip to content

Instantly share code, notes, and snippets.

@mdvsh
Created March 23, 2020 14:43
Show Gist options
  • Save mdvsh/b03f73b4bf8aaacab5bf919755b049a3 to your computer and use it in GitHub Desktop.
Save mdvsh/b03f73b4bf8aaacab5bf919755b049a3 to your computer and use it in GitHub Desktop.
Intel Scene Dataset Image Classification | Rank-1 98.2% Accuracy | fastAI
Display the source blob
Display the rendered blob
Raw
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@mdvsh
Copy link
Author

mdvsh commented Mar 23, 2020

Intel Scene Classification Challenge

Best so far: 98.18% accuracy achieved

Image classification of the intel-scene dataset leveraging the fastai v1. The goal is to try hit 95%+ accuracy, starting with a basic fastai image classification workflow and interating from there using many improvement stratergies.

This was all run on a GCP VM instance with NVIDIA P100 GPU

Check out the people I beat in this competition here

image

Setup

pseudocodenerd@sharmadhavs
-------------------------
OS: Ubuntu 18.04.4 LTS x86_64
Host: Inspiron 5567
Kernel: 5.3.0-42-generic
CPU: Intel i5-7200U (4) @ 3.100GHz
GPU: Intel HD Graphics 620
GPU: AMD Radeon R7 M260/M265
Memory: 8671MiB / 15911MiB

Google Cloud VM Instance
-------------------------
image-family="pytorch-latest-gpu"\
zone="asia-southt1-b"\
instance-name="fastai-madhav"\
instance-type="n1-highmem-8\
image-project=deeplearning-platform-release\
accelerator="type=nvidia-tesla-p100,count=1"\
boot-disk-size=200GB\
metadata="install-nvidia-driver=True"\

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment