Awesome

Image Captioning

We reimplemented the complicated Google' Image Captioning model by simple TensorLayer APIs.

This script run well under Python2 or 3 and TensorFlow 0.10 or 0.11.

1. Prepare MSCOCO data and Inception model

Before you run the scripts, you need to follow Google's setup guide, and setup the model, ckpt and data directories in *.py.

• Creat a data folder.
• Download and Preprocessing MSCOCO Data click here
• Download the Inception_V3 CKPT click here

2. Train the model

• via train.py

3. Evaluate the model

• via evaluate.py

4. Generate captions by given image and model

• via run_inference.py

5. Evaluation

• tylin/coco-caption