KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #dsh #dsh-plugin #ai-agent #codex #ai-agents #deepseek-harness #claude-code #cli #react #mcp

img2vec-keras

★ 37 Open GitHub ↗

Image to dense vector embedding. Clone of https://github.com/christiansafka/img2vec for Keras users

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# img2vec-keras
Image to dense vector embedding. This library uses the ResNet50 model in TensorFlow Keras, pre-trained on Imagenet, to generate image embeddings via https://keras.io/getting-started/faq/#how-can-i-obtain-the-output-of-an-intermediate-layer. Basically a clone of https://github.com/christiansafka/img2vec for TensorFlow Keras users.

# Install
`img2vec_keras` uses the `keras` module shipped with `tensorflow`. To install `img2vec_keras` and its dependencies

```pip install git+git://github.com/jaredwinick/img2vec-keras.git```

# Usage
```python
from img2vec_keras import Img2Vec
img2vec = Img2Vec()
x = img2vec.get_vec('/path/to/image/dog1.jpg')
```

# Examples

[Basic example with cosine similarity of vectors](https://github.com/jaredwinick/img2vec-keras/blob/master/examples/similarity.py)

[Colab notebook using t-SNE to visualize image vectors](https://colab.research.google.com/drive/14OvmH6KvoQJ41jb6QRL3FgwI61vq-UAJ)

# Contributors
* Thanks to @gmgeorg for upgrading to TensorFlow 2.0

img2vec-keras

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README