Multimodal-CoT incorporates vision features in a decoupled training framework. The framework consists of two training stages: (i) rationale generation and (ii) answer inference. Both stages share the same model architecture but differ in the input and output.
Install all required python dependencies:
pip install -r requirements.txt
Download the dataset from the following repository:
https://github.com/lupantech/ScienceQA/tree/main/data
The vision features (detr, resnet, clip, vit) are available at https://huggingface.co/cooelf/vision_features/tree/main
Alternatively, you may download the extracted vision features (detr, resnet, clip) from vision_features and unzip the files under vision_features
The processed vision features for ScienceQA are available at https://huggingface.co/cooelf/vision_features/tree/main.
The following instructions show how we obtain those features.
Download the image files from Google Drive and unzip all the images (train, dev, test) in the same folder (). The structure should be:
images
βββ 1
β βββ image.png
βββ 2
β βββ image.png
βββ 3
β βββ image.png
βββ 5
β βββ image.png
βββ 7
β βββ image.png
Run extract_features.py --data_root images --output_dir vision_features --img_type vit
If you hope to use your own images, please structure those images in the way above, or modify the script extract_features.py
.
The processed captions for ScienceQA are available at data/instruct_captions.json
.
The following instructions show how we obtain those features.
Intall lavis and prepare Vicuna weights to use InstructBLIP for caption extraction.
Assume that the images are stored in the images
folder.
python extract_caption.py
# rationale generation
CUDA_VISIBLE_DEVICES=0,1,2,3 python main.py \
--data_root data/ScienceQA/data \
--caption_file data/instruct_captions.json \
--model declare-lab/flan-alpaca-large \
--user_msg rationale --img_type vit \
--bs 2 --eval_bs 4 --epoch 50 --lr 5e-5 --output_len 512 \
--use_caption --use_generate --prompt_format QCM-E \
--output_dir experiments
# answer inference
CUDA_VISIBLE_DEVICES=0,1,2,3 python main_central.py \
--data_root data/ScienceQA/data \
--caption_file data/instruct_captions.json \
--model declare-lab/flan-alpaca-large \
--user_msg answer --img_type vit \
--bs 4 --eval_bs 8 --epoch 50 --lr 5e-5 --output_len 64 \
--use_caption --use_generate --prompt_format QCMG-A \
--output_dir experiments \
--eval_le experiments/rationale_declare-lab-flan-alpaca-large_vit_QCM-E_lr5e-05_bs8_op512_ep50/predictions_ans_eval.json \
--test_le experiments/rationale_declare-lab-flan-alpaca-large_vit_QCM-E_lr5e-05_bs8_op512_ep50/predictions_ans_test.json
Our trained models are available at https://huggingface.co/cooelf/mm-cot/tree/main. To use our trained models, please put the them under the models
folder.
# rationale generation
CUDA_VISIBLE_DEVICES=0,1,2,3 python main.py \
--data_root data/ScienceQA/data \
--caption_file data/instruct_captions.json \
--model declare-lab/flan-alpaca-large \
--user_msg rationale --img_type vit \
--bs 2 --eval_bs 4 --epoch 50 --lr 5e-5 --output_len 512 \
--use_caption --use_generate --prompt_format QCM-E \
--output_dir experiments
--evaluate_dir models/mm-cot-large-rationale
# answer inference
CUDA_VISIBLE_DEVICES=0,1,2,3 python main_central.py \
--data_root data/ScienceQA/data \
--caption_file data/instruct_captions.json \
--model declare-lab/flan-alpaca-large \
--user_msg answer --img_type vit \
--bs 4 --eval_bs 8 --epoch 50 --lr 5e-5 --output_len 64 \
--use_caption --use_generate --prompt_format QCMG-A \
--output_dir experiments \
--eval_le experiments/rationale_declare-lab-flan-alpaca-large_vit_QCM-E_lr5e-05_bs8_op512_ep50/predictions_ans_eval.json \
--test_le experiments/rationale_declare-lab-flan-alpaca-large_vit_QCM-E_lr5e-05_bs8_op512_ep50/predictions_ans_test.json \
--evaluate_dir models/mm-cot-large-answer
@article{zhang2023multicot,
title={Multimodal Chain-of-Thought Reasoning in Language Models},
author={Zhang, Zhuosheng and Zhang, Aston and Li, Mu and Zhao, Hai and Karypis, George and Smola, Alex},
journal={arXiv preprint arXiv:2302.00923},
year={2023}
}
This project is licensed under the Apache-2.0 License.
Part of our codes are adapted from ScienceQA, Transformers, pytorch-image-models.
We thank Pan Lu for providing parameter size for ScienceQA baselines.