https://github.com/quanpersie2001/imagecaptioning
Predicting a caption for a given image using Inception Net V3, LSTM and Glove
https://github.com/quanpersie2001/imagecaptioning
cnn image-captioning imagecaptioning inception-v3 lstm rnn
Last synced: about 2 months ago
JSON representation
Predicting a caption for a given image using Inception Net V3, LSTM and Glove
- Host: GitHub
- URL: https://github.com/quanpersie2001/imagecaptioning
- Owner: quanpersie2001
- Created: 2022-12-13T05:40:41.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-05-20T19:35:34.000Z (about 2 years ago)
- Last Synced: 2025-02-15T02:43:20.570Z (3 months ago)
- Topics: cnn, image-captioning, imagecaptioning, inception-v3, lstm, rnn
- Language: Jupyter Notebook
- Homepage:
- Size: 14.4 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Image Captioning
## Model using Inception Net V3, LSTM and Glove (Using SSD300 to improve feature)
## How to run?
### Install lib
```
pip install -r requirements.txt
```
### Download data
```console
python data_download.py
```
> **Note** : Dataset is MS COCO 2014 and Glove . This is large dataset, long download.
### Preprocess
You **must** run
```console
python preprocess.py
```
With COCO datase this command runs for a long time you can download and coppy them to `ROOT / process_data`
### [Download here](https://drive.google.com/drive/folders/1HDgToaiFKzVNTQZI1ts2Dlfgh1sVMk3D?usp=sharing)
### Trainning
```console
python train.py --batch-size 64 --output weights --epochs 30
```
You can download pre-train model an copy them to `ROOT / weights`
### [Download here](https://drive.google.com/drive/folders/1oXVC8fVioblaRpvB-tVtQsBHwKTmfMse?usp=sharing)
### Predict
```console
python predict.py --image path/to/image --weight path/to/weight --k-beam 9
```
## Result
You can see sumary in [summary.ipynb](summary.ipynb)