Skip to main content

Restore the official code 100% and improve it to make it easier to use.

Project description


Update (Feb 17, 2020)

The update is for ease of use and deployment.

It is also now incredibly simple to load a pretrained model with a new number of classes for transfer learning:

from googlenet_pytorch import GoogLeNet 
model = GoogLeNet.from_pretrained('googlenet')


This repository contains an op-for-op PyTorch reimplementation of Going Deeper with Convolutions.

The goal of this implementation is to be simple, highly extensible, and easy to integrate into your own projects. This implementation is a work in progress -- new features are currently being implemented.

At the moment, you can easily:

  • Load pretrained GoogLeNet models
  • Use VGGNet models for classification or feature extraction

Upcoming features: In the next few days, you will be able to:

  • Quickly finetune an GoogLeNet on your own dataset
  • Export GoogLeNet models for production

Table of contents

  1. About GoogLeNet
  2. Installation
  3. Usage
  4. Contributing

About GoogLeNet

If you're new to GoogLeNet, here is an explanation straight from the official PyTorch implementation:

We propose a deep convolutional neural network architecture codenamed "Inception", which was responsible for setting the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC 2014). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. This was achieved by a carefully crafted design that allows for increasing the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC 2014 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.


Install from pypi:

pip install googlenet_pytorch

Install from source:

git clone
cd GoogLeNet-PyTorch
pip install -e .


Loading pretrained models

Load a pretrained GoogLeNet:

from googlenet_pytorch import GoogLeNet
model = GoogLeNet.from_pretrained("googlenet")

Their 1-crop error rates on imagenet dataset with pretrained models are listed below.

Model structure Top-1 error Top-5 error
googlenet 30.22 10.47

Example: Classification

We assume that in your current directory, there is a img.jpg file and a labels_map.txt file (ImageNet class names). These are both included in examples/simple.

All pre-trained models expect input images normalized in the same way, i.e. mini-batches of 3-channel RGB images of shape (3 x H x W), where H and W are expected to be at least 224. The images have to be loaded in to a range of [0, 1] and then normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225].

Here's a sample execution.

import json

import torch
import torchvision.transforms as transforms
from PIL import Image

from googlenet_pytorch import GoogLeNet 

# Open image
input_image ="img.jpg")

# Preprocess image
preprocess = transforms.Compose([
  transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
input_tensor = preprocess(input_image)
input_batch = input_tensor.unsqueeze(0)  # create a mini-batch as expected by the model

# Load class names
labels_map = json.load(open("labels_map.txt"))
labels_map = [labels_map[str(i)] for i in range(1000)]

# Classify with GoogLeNet
model = GoogLeNet.from_pretrained("googlenet")

# move the input and model to GPU for speed if available
if torch.cuda.is_available():
  input_batch ="cuda")"cuda")

with torch.no_grad():
  logits = model(input_batch)
preds = torch.topk(logits, k=5).indices.squeeze(0).tolist()

for idx in preds:
  label = labels_map[idx]
  prob = torch.softmax(logits, dim=1)[0, idx].item()
  print(f"{label:<75} ({prob * 100:.2f}%)")

Example: Feature Extraction

You can easily extract features with model.extract_features:

import torch
from googlenet_pytorch import GoogLeNet 
model = GoogLeNet.from_pretrained('googlenet')

# ... image preprocessing as in the classification example ...
inputs = torch.randn(1, 3, 224, 224)
print(inputs.shape) # torch.Size([1, 3, 224, 224])

features = model.extract_features(inputs)
print(features.shape) # torch.Size([1, 1024, 7, 7])

Example: Export to ONNX

Exporting to ONNX for deploying to production is now simple:

import torch 
from googlenet_pytorch import GoogLeNet 

model = GoogLeNet.from_pretrained('googlenet')
dummy_input = torch.randn(16, 3, 224, 224)

torch.onnx.export(model, dummy_input, "demo.onnx", verbose=True)

Example: Visual

cd $REPO$/framework

Then open the browser and type in the browser address

Enjoy it.


See examples/imagenet for details about evaluating on ImageNet.

For more datasets result. Please see research/


If you find a bug, create a GitHub issue, or even better, submit a pull request. Similarly, if you have questions, simply post them as GitHub issues.

I look forward to seeing what the community does with these models!

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for googlenet-pytorch, version 0.1.7
Filename, size File type Python version Upload date Hashes
Filename, size googlenet_pytorch-0.1.7-py2.py3-none-any.whl (13.7 kB) File type Wheel Python version py2.py3 Upload date Hashes View
Filename, size googlenet_pytorch-0.1.7.tar.gz (10.2 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page