Self-critical Sequence Training for Image Captioning (+ misc.)

This repository includes the unofficial implementation Self-critical Sequence Training for Image Captioning and Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering.

The author of SCST helped me a lot when I tried to replicate the result. Great thanks. The att2in2 model can achieve more than 1.20 Cider score on Karpathy's test split (with self-critical training, bottom-up feature, large rnn hidden size, without ensemble)

This is based on my ImageCaptioning.pytorch repository. The modifications is:

Self critical training.
Bottom up feature support from ref. (Evaluation on arbitrary images is not supported.)
Ensemble
Multi-GPU training
Add transformer (merged from Transformer_captioning)

Requirements

Python 2 or 3 (My coco-caption supports python 3)

PyTorch 1.3 (along with torchvision)

cider (already been added as a submodule)

coco-caption (already been added as a submodule) (Remember to follow initialization steps in coco-caption/README.md)

yacs

(Skip if you are using bottom-up feature): If you want to use resnet to extract image features, you need to download pretrained resnet model for both training and evaluation. The models can be downloaded from here, and should be placed in data/imagenet_weights.

Pretrained models

Checkout MODEL_ZOO.md.

If you want to do evaluation only, you can then follow this section after downloading the pretrained models (and also the pretrained resnet101 or precomputed bottomup features).

Train your own network on COCO/Flickr30k

Prepare data.

We now support both flickr30k and COCO. See details in data/README.md. (Note: the later sections assume COCO dataset; it should be trivial to use flickr30k.)

Start training

$ python train.py --id fc --caption_model newfc --input_json data/cocotalk.json --input_fc_dir data/cocotalk_fc --input_att_dir data/cocotalk_att --input_label_h5 data/cocotalk_label.h5 --batch_size 10 --learning_rate 5e-4 --learning_rate_decay_start 0 --scheduled_sampling_start 0 --checkpoint_path log_fc --save_checkpoint_every 6000 --val_images_use 5000 --max_epochs 30

$ python train.py --cfg configs/fc.yml --id fc

The train script will dump checkpoints into the folder specified by --checkpoint_path (default = log_$id/). By default only save the best-performing checkpoint on validation and the latest checkpoint to save disk space. You can also set --save_history_ckpt to 1 to save every checkpoint.

To resume training, you can specify --start_from option to be the path saving infos.pkl and model.pth (usually you could just set --start_from and --checkpoint_path to be the same).

To checkout the training curve or validation curve, you can use tensorboard. The loss histories are automatically dumped into --checkpoint_path.

The current command use scheduled sampling, you can also set --scheduled_sampling_start to -1 to turn off scheduled sampling.

If you'd like to evaluate BLEU/METEOR/CIDEr scores during training in addition to validation cross entropy loss, use --language_eval 1 option, but don't forget to pull the submodule coco-caption.

For all the arguments, you can specify them in a yaml file and use --cfg to use the configurations in that yaml file. The configurations in command line will overwrite cfg file if there are conflicts.

For more options, see opts.py.

Train using self critical

First you should preprocess the dataset and get the cache for calculating cider score:

$ python scripts/prepro_ngrams.py --input_json data/dataset_coco.json --dict_json data/cocotalk.json --output_pkl data/coco-train --split train

Then, copy the model from the pretrained model using cross entropy. (It's not mandatory to copy the model, just for back-up)

$ bash scripts/copy_model.sh fc fc_rl

Then

$ python train.py --id fc_rl --caption_model newfc --input_json data/cocotalk.json --input_fc_dir data/cocotalk_fc --input_att_dir data/cocotalk_att --input_label_h5 data/cocotalk_label.h5 --batch_size 10 --learning_rate 5e-5 --start_from log_fc_rl --checkpoint_path log_fc_rl --save_checkpoint_every 6000 --language_eval 1 --val_images_use 5000 --self_critical_after 30 --cached_tokens coco-train-idxs --max_epoch 50 --train_sample_n 5

$ python train.py --cfg configs/fc_rl.yml --id fc_rl

You will see a huge boost on Cider score, : ).

A few notes on training. Starting self-critical training after 30 epochs, the CIDEr score goes up to 1.05 after 600k iterations (including the 30 epochs pertraining).

Generate image captions

Evaluate on raw images

Note: this doesn't work for models trained with bottomup feature. Now place all your images of interest into a folder, e.g. blah, and run the eval script:

$ python eval.py --model model.pth --infos_path infos.pkl --image_folder blah --num_images 10

This tells the eval script to run up to 10 images from the given folder. If you have a big GPU you can speed up the evaluation by increasing batch_size. Use --num_images -1 to process all images. The eval script will create an vis.json file inside the vis folder, which can then be visualized with the provided HTML interface:

$ cd vis
$ python -m SimpleHTTPServer

Now visit localhost:8000 in your browser and you should see your predicted captions.

Evaluate on Karpathy's test split

$ python eval.py --dump_images 0 --num_images 5000 --model model.pth --infos_path infos.pkl --language_eval 1

The defualt split to evaluate is test. The default inference method is greedy decoding (--sample_method greedy), to sample from the posterior, set --sample_method sample.

Beam Search. Beam search can increase the performance of the search for greedy decoding sequence by ~5%. However, this is a little more expensive. To turn on the beam search, use --beam_size N, N should be greater than 1.

Evaluate on COCO test set

$ python eval.py --input_json cocotest.json --input_fc_dir data/cocotest_bu_fc --input_att_dir data/cocotest_bu_att --input_label_h5 none --num_images -1 --model model.pth --infos_path infos.pkl --language_eval 0

You can download the preprocessed file cocotest.json, cocotest_bu_att and cocotest_bu_fc from link.

Miscellanea

Using cpu. The code is currently defaultly using gpu; there is even no option for switching. If someone highly needs a cpu model, please open an issue; I can potentially create a cpu checkpoint and modify the eval.py to run the model on cpu. However, there's no point using cpus to train the model.

Train on other dataset. It should be trivial to port if you can create a file like dataset_coco.json for your own dataset.

Live demo. Not supported now. Welcome pull request.

For more advanced features:

Checkout ADVANCED.md.

Reference

If you find this repo useful, please consider citing (no obligation at all):

@article{luo2018discriminability,
  title={Discriminability objective for training descriptive captions},
  author={Luo, Ruotian and Price, Brian and Cohen, Scott and Shakhnarovich, Gregory},
  journal={arXiv preprint arXiv:1803.04376},
  year={2018}
}

Of course, please cite the original paper of models you are using (You can find references in the model files).

Acknowledgements

Thanks the original neuraltalk2 and awesome PyTorch team.

wouterkool/self-critical.pytorch