chuanyangjin / mmtom-qa Goto Github PK

Hi Chuanyang,

Could you release the inference_raw.json file used in text_parsing.py for test evaluation?

To reproduce the _BIP-ALM_method test, we need the multimodal_representations.json file. However, generating this file requires running text_parsing.py with inference_raw.json, which is not currently available in the repo (and also lacks instructions on how to create it).

If the inference_raw.json file can be released (or, more straightforwardly, if the final evaluation test data multimodal_representations.json for BIP-ALM can be released), it will enable successfully reproducing the test evaluation of the proposed BIP-ALM method. This would help to enhance the impact and reproducibility of this smart paper within the community.

Thank you very much!

Could you release the TEST_INDICES and finetuning_data used in graph_utils.py and generate_finetuning_dataset.py?

Thanks!

Hi Chuanyang,

Could you release the extracted multimodal_representations.json from text_parsing.py?

I found it would be directly useful for reproducing the claimed inference results of BIP-ALM reported in the MMToM-QA paper, and it seems that this .json file is not currently released in the repo for test evaluation.

Thank you very much!

P.S. If the extracted .json files from python visual_perception.py --benchmark_video_path videos could also be released, I would highly appreciate it.